Skip to main content

Full text of "The_Chinese_Kinship_System_"

See other formats





Cambridge, Massachusetts 



Reprinted from the Harvard Journal of Asiatic Studies 
Vol. II, No. 2 (July, 1937), pp. 141ff. 

Second Printing 

Distributed in Great Britain by Oxford 
University Press, London 

Printed in the United States of America 


Ever since its first publication in the Harvard Journal of 
Asiatic Studies, Volume II (pp. 141-275), 1937, Feng Han-yi's 
article has remained one of the basic works on the Chinese kin- 
ship system. It was separately issued by the Harvard University 
Press in 1948, with no change other than repagination. In re- 
sponse to continuing demand, this version is again reprinted. 
For convenience of distribution it is included in the Harvard- 
Yenching Institute Series, which is financed from the residue of 
funds granted by the Rockefeller Foundation during World War 
II for the publication of Chinese and Japanese dictionaries. 




Abbreviations % 

Definition of terms 2 

Chronology 2 

Introduction 4 

Principles of Terminological Composition .... 8 

Nuclear Terms 8 

Basic Modifiers 10 

Terminological Composition 11 

Referential Modifiers 13 

Vocative Terms 17 

Supernumerary Terms 19 

Structural Principles and Terminological Categories 20 

Lineal and Collateral Differentiation 20 

Generation Stratification 24 

Categories 27 

Reciprocity SO 

Factors Affecting the System 30 

The Sib: Decent and Exogamy 33 

Mourning Grades 38 

Cross-cousin Marriage 43 

Sororate 40 

Levirate 51 

Teknonymy 54 

Historical Review of Terms 04 

Consanguineal Relatives: 67 

Relatives through father — Table I 67 

Relatives through mother — Table II ... 107 

Affinal Relatives .... 114 

Relatives through wife — Table III 114 

Relatives through husband — Table IV 121 

Conclusions \%{\ 

Chinese Works Frequently Cited 130 



e ~ ego h a husband, husband's 

f = father, father's w = wife, wife's 

m = mother, mother's o = older 

s s son, son's y = younger 

d = daughter, daughter's > = older than 

b - brother, brother's < = younger than 
si = sister, sister's 

Example: m f b s d > e signifies mother's father's brother's son's daughter older than ego. 

Definition of Terms 

All terms are used in their customary meanings, as found in anthropological and 
sociological literature. A few terms are used here with a specialized connotation in 
connection with the Chinese social system. They are the following: 

Family: used always in the sense of the " extended family " or the Gross-Familie, and 
equivalent to the Chinese term chia |§C> or chia ting ^tj£|. 

Sib: a group of people possessing a common sibname (patronym), descended from a 
common male ancestor, no matter how remote, and characterized by a feeling of 
relationship. Descent is strictly patrilineal, and the group is strictly exogamous. 
An organization for the common welfare of all its members, and ancestor 
u worship," may or may not be present. It is equivalent to the Chinese term, 
tsung tsu ^$e> 

Sibname: used in the sense of a patronym or surname which all members of a sib 
possess in common, and equivalent to the Chinese term hsing |ft. Descent of 
the sibname is strictly patrilineal. 

Sib relative: relatives who belong to the same sib and possess the same sibname as 
ego. It is equivalent to the Chinese term tsung ch'in ^;|S? , or tsu jen )}%J^ 
'* clansmen." 

Non-sib relative: relative who belongs to a sib other than ego's and bears a sibname 
other than ego's. It is equivalent to the Chinese terms wai ch'in ^f*^ and net 
ch'in F^Jl! combined; or the old legal term ch'in shu KtliJ. Wm ch'in refers 
to relatives through women of the sib married out, and the affinal relatives of 
father, fathers father, and ascending. Nei ch'in refers to ego's own affinal 


The following chronology is given for those who are not familiar with Chinese history 
since it is impossible to give the Western date in every instance. The tripartite division 
does not correspond to the traditional Chinese historical divisions but has been adopted 
here simply with reference to the evolution of the kinship system. 

Chronology 3 

Ancient period: first millennium B. C, which includes the following dynastic periods: 

Chou dynasty, or the feudal period, ca. 1100-249 B.C. 
Ch'in dynasty, 248-207 B.C. 

Former Han dynasty, 206 B. C.-24 A. D. 

Transitional period: first millennium A. D., which includes the following dynastic periods: 

Later Han dynasty, 25-220 A. D. 

Wei dynasty, 220-264 A.D. 

Chin dynasty, 265-420 A.D. 

Sung dynasty, 420-479 A. D Northern Wei dynasty, 399-534 A. D. 

Ch'i dynasty, 479-502 A. D. Northern Ch'i dynasty, 550-577 A. D. 

Liang dynasty, 502-557 A. D. 

Ch'en dynasty, 557-589 A.D. 

Sui dynasty, 581-618 A.D. 

Tang dynasty, 618-907 A. D. 

Wu tai, 907-960 A. D. 

Modern period: second millennium A. D., which includes the following dynastic periods: 

Sung dynasty, 960-1279 A. D. (This is the Sung dynasty to which the 

Yuan dynasty, 1280-1368 A. D. writer will always refer in the present 

Ming dynasty, 1368-1644 A. D. treatise, not to the one mentioned above 

Ch'ing dynasty, 1644-1911 A.D. under 420-479 A.D.) 

[Other contemporary dynastic periods are omitted here since they are not 
referred to in this work.] 

Ancient system: system of the ancient period, i. e., the system in the Erh Ya, supple- 
mented by the / Li, the Li Chi, and other contemporary sources. 
Modern system: system of the modern period, i. e., the present Chinese system. 


The interest of the Chinese themselves in problems of kinship 
was manifested quite early. This interest is primarily a practical 
one, for the whole Chinese social structure is built upon the basis 
of the " extended family " organization, which in turn is based 
upon the systematization of the mutual relationships among its 
members. If the whole social structure is to function harmoni- 
ously, the kinship system, which expresses and defines the rights 
and obligations of individuals to each other, must first be ad- 
justed. This ideology is further fostered by the teachings of 
Confucianism, so that kinship becomes a subject of perennial 

The systematic recording of relationship terms goes back as 
far as the Erh Ya, 1 a work of the third or second century B. C, 
(according to the more conservative dating) , in which the terms 
are carefully classified and arranged. Subsequent works of a 
similar nature all contain special chapters on kinship nomencla- 
ture, e. g., the Shih Ming [ca. 200 A. D.] and the Kuang Ya [ca. 
230 A. D.] — to mention only two of the comparatively earlier 
ones. These works record later terms which are not present in 
the Erh Ya and in a sense bring the Erh Ya system up to the 
date of each compilation. This practice has continued down to 
the present day. 2 Even larger encyclopaedic works devote special 
sections to this subject, e. g., the Tax ping yu Ian (983 A. D.) 

tradition has it that the ErhYa was compiled by Chou Ktmg [P-1105 B.C.] and 
augmented by Confucius [551-479 B.C.], Tzu Hsia -^H [S07-? B.C.], Shu-sun Tung 
^Uftpfi [° a * 200 B. C.J and others. It is not the work of one hand, nor of one period, 
but developed gradually during the first millennium B. C. Cf. B. Karlgren, BMFEA 
8(1931). 44-49. The section on Kinship Terms f^H probably dates from ca. 200 
B.C. Cf. mm%m$l by Naito Torajiro, in ;^Sf£# 2.163-184. 

2 Most works of the Ullfjfe class nave a section on kinship terms, e. g., the P'ien ya 
Uf^fl [under 5. f^|Jjf ], Shih yaffil , etc., and even works on dialects like the 
Hsu fang yen ^^fW of Hang Shih-chiin $i -ffr|g [1696-1773], and the I yu Mf§ 
of Ch'ien Tien j$£tij [1744-1806] devote special sections to kinship term variants. 
Other dictionaries, starting from the Skuo wen, contain kinship terms too, but they 
are not systematically arranged. 


has ten chapters, 3 and the Tu shu chi ch'eng 112 chapters, 4 on 
kinship nomenclature. Naturally, not all this material is relevant, 
and much of it belongs to belles lettres. In the Ch'ing dynasty a 
series of special works on kinship terms appeared, the most im- 
portant and extensive of which are the Ch'eng wei lu fflMM of 
Liang Chang-chii ^IB [1775-1849], and the CKin shu chi M 
Mti of Cheng Chen M& [1806-1864]. Both of these works com- 
prise collections of old terms, and are more or less in the Erh Ya 
tradition. Of the two, the CKeng wei lu is much wider in scope, 
but the arrangement of material is rather loose. The CKin shu 
chi considers only the lineal relatives through father, and the 
ordering of data is more in keeping with the view of orthodox 
Confucianism. 5 

By far the most important class of materials is formed by the 
ritual works, the Li. In these works kinship is not treated as a 
subject by itself but in connection with other subjects ; an excep- 
tion is the CKin shu chi, a lost section of the Li, which dealt pri- 
marily with relationship terminology. 6 These ritual works are 
important sources for the functional study of the Chinese kin- 
ship system because they deal with kinship in action. Such are the 
/ Li and the Li Chi, works of the second half of the first millenium 
B. C, that treat kinship in extenso, especially in connection with 
mourning rites, ancestor " worship," and other aspects of ritual. 
In all later works on ritualism — too numerous to mention here — 
kinship is the basic subject of discussion. 

In addition, there are numerous miscellaneous works in which 
discussions on kinship terms occasionally occur. These are among 
the most important sources from the standpoint of the evolution 
of the Chinese kinship system, because generally it is here that 

4 ■««$, wmma- sue*, i-m.. 

B Other important works are the Tung su pien ^tf&lg [ch. 4: Hfrfft, ch. 18: f|$ji|] 
ofCHAiHao gffl.?-lTO8; the Heng yen lu ®g^ [ch. S: SJKfiHR&l of Ch'ien 
Ta-hsin ^^cBJf , 1727-1804; the Cheng su wen Wifc% tck. 4] of Ho I-hsing Sfktl 
1757-1825; and the Kuang shih ch'in J^rf fg of Chang Shen-i 36® |§. There are 
many other works of a less extensive nature but they rather duplicate each other. 

8 As quoted in Pai hu t'ung 8. 19b, the CKin shu chi fijgf £ is very similar in nature 
to the Erh Ya. 


one finds recorded the newly introduced terms (dialectical or 
unconventional) which, as a rule, are ignored by the ritual and 
other formal literature. Very often one finds in them enlightening 
discussions concerning the introduction and origin of new 

Thus, the interest in the study of kinship terminology is not 
new among Chinese scholars, and actually they sometimes made 
explanations which might rank with modern sociological inter- 
pretations, but the systematic socio-anthropological study began 
with Lewis H. Morgan. 7 Morgan's data were supplied by Robert 
Hart, an Englishman in the employ of the Chinese Maritime 
Customs. Despite the faulty nature of Hart's material, and not- 
withstanding Morgan's evolutionistic predilections, which invali- 
dated most of his conclusions, the Morgan-Hart work has re- 
mained the basis for most subsequent speculation. Since then has 
appeared a number of miscellaneous recordings, some in legal 
treatises or linguistic primers, others in lexicographic works, but, 
with one exception, none is worthy of serious consideration. This 
exception is the work of T. S. Chen and J. K. Shryock. 8 The 
Chen-Shryock study is based chiefly on two modern dictionaries, 
the Tzu yuan and the Chung hua ta tzu lien. This material, 
although inadequate and unreliable, has been used to good ad- 
vantage by the authors. George W. Bounakoff 9 seems to have 
made a stupendous attempt to synthesize all the material in 
European languages in the light of Morgan's hypothesis. 

7 Systems of Consanguinity and Affinity of the Human Family, 1870, Part III, ch. 
IV. 413-437. 

8 " Chinese Relationship Terms," American Anthropologist 34 (1932) . 623-664. 

8 Terms of Relationship in Chinese: An Ethnographical-Linguistic Study [N. J. Marr 
Institute of Language and Mentality, Academy of Sciences, USSR, 1936]. I have not 
seen the original work (in Russian) but only the English " Analytical Summary." So 
far as I can make out, it is mainly based on secondary English sources with the 
exception of Erh Ya, which is also available in English. Although he has used most 
of the European sources, he seems to have overlooked the tables of Gustave Schlegel: 
Nederlandsch-Chineesch woordenboek met de transcriptie der Chineesche karackters in 
het Tsiang-tsiu dialekt, Leiden, 1886-90, vol. I, p. 1343, Chineesche Geslachtboon. 
The important issue, however, is his methodology which is based upon the Marxian 
conception of history and the Japhetic theory of language of H. I. Marr. Combining 
these with the evolutionary stages of L. H. Morgan, he arrives at the M collective 
beginnings " of Chinese society ! 


The limitations of all these European works are obvious. First, 
the terms relied upon have been collected by untrained persons 
from " uninformed " informants. Secondly, the writers have not 
made use of the vast amount of easily available Chinese docu- 
mentary material — indeed, they seem unaware of the existence 
of such material. Consequently, most of these studies are marred 
by numerous inconsistencies and errors. The writers could seldom 
determine the exact nature of a term, because few ever sectioned 
out the multiple strata of terminology in the Chinese system. 10 

The present study is based chiefly upon the author's own col- 
lection of terms from the primary Chinese sources. All material 
consulted has been examined critically to insure correctness in 
terminology and interpretation. The method of approach is pri- 
marily historical and linguistic 11 — partly because it is precisely 
these aspects which are the most engaging characteristics of the 
Chinese system and at the same time the least understood, partly 
because preliminary work of this sort is prerequisite to an under- 
standing of the more implicit aspects of the system. The field of 
investigation is limited to the historical period, approximately the 
last twenty-five centuries, within which the system has been fully 
documented. If kinship system changes occur at all, two thousand 
years should be long enough for their manifestation. 12 

10 Originally the present work included a section called "Critical Review of Early 
Studies." As it was not in a way constructive and occupied considerable space, it has 
been deleted. 

11 By linguistic, I mean the usual and the formal approach to the study of kinship 
systems, i. e., by an examination of the kinship terms themselves, the structural whole 
they present, and the underlying principles involved. 

12 This paper is one of the studies made by the author under a fellowship grant 
by the Trustees of the Harvard- Yenching Institute, to whom he wishes to express his 
gratitude. The author is also deeply indebted to Professors A. I. Halloweli, F. G. 
Speck, and D. S. Davidson for constant guidance, suggestions and improvements in 
this work; to Prof. C. K. M. Kiuckhohn and Dr. C. M. Arensberg for their careful 
examination and corrections of the MS.; to Mr. Paul K. Benedict, who has made 
improvements and corrections on almost every page. The author also wishes to thank 
Professors S. Elisseeff and J. R. Ware for suggestions and assistance and especially 
for sponsoring publication in this Journal. 


The principles governing the composition of terms are both 
linguistic and sociological. Linguistically, they are formed accord- 
ing to the syntactical principles of the Chinese language; socio- 
logically, their connotations are determined by the relationships 
which they express and the circumstances under which they are 
used. The multitude of Chinese relationship terms can be reduced 
to four fundamental classes, namely, nuclear terms, basic modifiers, 
referential modifiers, and vocatives. Nuclear terms express the 
nuclear group of relationships and, linguistically, are independent 
of modifiers. Each nuclear term possesses a primary meaning 
and one or more secondary meanings. The primary meaning is 
assumed when the term is used independently, and, when it is 
used in combination with other elements, the secondary meaning 
or meanings become paramount. Basic modifiers for the most 
part express collateral relationship and generation status and can- 
not be used independently as kinship terms. The nuclear terms 
form the basis for kinship extensions and the basic modifiers 
locate the exact place of the relative in the total scheme. The 
combinations and recombinations of these two classes of elements 
constitute the modern standard system which is the norm of all 
other terminologies. The referential modifiers modify the standard 
system into proper forms for referential use in specific applications. 
Vocatives, aside from their primary usages, transform them into 
direct forms of address between relatives. 

The following is an analysis of these four classes of elements 
and an exposition of the principles governing their composition 
and application. In the analysis of the nuclear terms the primary 
connotations (according to the modern standard system) are 
given first and then followed by the secondary meanings. 

Nuclear Terms 

Tsu jjj& : 3 Father's father. Ancestor. Used in combination with other elements for all 

ascendants higher than the father's generation. 
Sun ^: 8 Son's son. Descendant. Used in combination with other elements for all 

descendants lower than the son's generation. 



Fu JC' 4, Father. Male of higher generation status. Male sex indicator for higher 
generations. Suffixed to terms of all male relatives of generations higher than 

Tzu-j-: B Son. Male of lower generation status. Male sex indicator for lower genera- 
tions. May be suffixed to terms of male relatives of generations lower than ego, 
but its use is optional. 

Mu -ft: 8 Mother. Female of higher generation status. Female sex indicator for higher 
generations. Suffixed to terms of all married female relatives of generations 
higher than ego. 

Nii ^: 7 Daughter. Female of lower generation status. Female sex indicator for lower 
generations. Suffixed to terms of all female relatives of generations lower than 

Hsiung fa: 8 Older brother. Male of the older brother's status. Indicator of seniority 
within the generation of ego. Used in combination with other elements for male 
relatives of the generation of, but older than, ego. 

Ti ifJ: B Younger brother. Male of younger brother's status. Indicator of juniority 
for males within the generation of ego. Used in combination with other elements 
for male relatives of the generation of, but younger than, ego. 

Tzu J$: 10 Older sister. Female of the older sister's status. Indicator of seniority for 
females within the same generation of ego. Used in combination with other 
elements for female relatives of the generation of, but older than, ego. 

Mei jfc: 11 Younger sister. Female of the younger sister's status. Indicator of juniority 
for females within the same generation of ego. Used in combination with other 
elements for female relatives of the generation of, but younger than, ego. 

Pofg: 12 Father's older brother. Indicator of seniority. Applicable to terms from ego's 
generation and ascending, using the direct male lineal line as a standard of 
Husband's older brother. Cannot be extended in this sense. 

Shu^: 13 Father's younger brother. Indicator of juniority. Applicable to terms from 
ego's generation and ascending, using the direct lineal line as a standard of 
Husband's younger brother. Cannot be extended in this last sense. 

Chih££: 13a Brother's son. Indicator of descent from male collaterals. Used in com- 
bination with other elements for descendants of male relatives of the generation 
of ego. 

Sheng JJg: 1 * Sister's son. Indicator of descent from female collaterals. Used in com- 
bination with other elements for descendants from female relatives of the 
generation of ego. 

Ku M- 15 Father's sister. Indicator of relationship comparable with father's sister's. 
Indicator of descent from father's sister, or from female relatives comparable 
with father's sister's relationship. 
Husband's sister. Cannot be extended in this last sense. 
Chiu J|: lfl Mother's brother. Indicator of relationship comparable with mother's 
brother's. Indicator of descent from mother's brother, or from male relatives 
comparable with mother's brother's relationship. 
Wife's brother. Cannot be extended in this last sense. 


I $|: Mother's sister. Indicator of relationship comparable with mother's sister's. 
Indicator of descent from mother's sister, or from female relatives comparable 
with mother's sister's relationship. Wife's sister. 17 Indicator of relationship 
comparable with wife's sister's. Indicator of descent from wife's sister, or from 
female relatives comparable with wife's sister's relationship. 

Yo-Jff: 17 * Wife's parents. Indicator of relationship comparable with wife's parents, 
such as their cousins. 

Hsu iff: 18 Daughter's husband. Husband. Indicator of connection by marriage with 
ego's female relatives of the same generation of ego and descending. 

Fu ffe: 1 * ego's husband. Husband. Indicator of connection by marriage with ego's 
female relatives of the same generation of ego. 

Ch'i|£: ao ego's wife. Wife. 

Sao^S?: 21 Older brother's wife. Female of older brother's wife's status. Indicator of 
connection by marriage with ego's male relatives of the generation of, but older 
than, ego. 

Fu $ff : 23 Son's wife. Wife. Indicator of connection by marriage with ego's male rela- 
tives of the generation of, but younger than, ego and descending generations. 

Basic Modifiers 

Kao jij: High; revered. Modifying indicator of the fourth ascending generation. 

Tseng 'H*: Added; increased. Modifying indicator for the third ascending and descend- 
ing generations. 

Hsiian j£: Far; distant. Modifying indicator for the fourth descending generation. 

Tang ^: Hall; the ancestral hall. Modifying indicator for the second collateral line 
from the second generation and descending. Ascending vertically, for father's 
father's brother's children and father's father's father's brother's children, that 
is, for paternal uncles and aunts once removed and paternal granduncles and 
aunts once removed. When extended to non-sib relatives, it indicates, in a 
similar way, the third collateral line. 

Ts'ung $£: To follow; through. It is used synonymously with fang. Tang is a later 
term and its use is restricted. Wherever fang is used, ts'ung may be substituted, 
but not vice versa. 

Tsai-tsung 4?$£: To follow again, or to follow a second time. Modifying indicator 
for the third collateral line from ego's generation and descending. Ascendingly, 
for paternal uncles and aunts twice removed. 

Tsu $£: Sib; tribe. Modifying indicator for relationships from the fourth collateral 
line and beyond. 

Piao ^$t\ Outside; external. Indicator of descent from father's sister, mother's brother 
and mother's sister. Similarly extended to all relatives descended from those 
whose terms include either ku (father's sister), chin (mother's brother) or i 
(mother's sister). 

Nei j^J: Inside; inner; wife. Indicator of descent from wife's brother, or from relatives 
comparable with him, e. g.» his male sib-cousins. 


Wai ^f. : Outside. Reciprocal modifier indicating mother's parents and daughter's 

The above generalizations are based on the connotations of the modern standard 
terminology. They are abstracted from the whole range of the nomenclature, with 
every term taken into consideration. Yet, because of the multitude of possible com- 
binations for every Chinese character, exceptions are inevitable. These exceptions, 
few and relatively insignificant, will be evident when the whole system is reviewed. 

Terminological Composition 

In the building of terms, the terminology for the nuclear group 
of relations is taken as a structural basis, with the exception of 
parent-child and hiilband-wife terms, which are used as sex indi- 
cators. 23 All modifying elements indicating collateral relationship 
and descent are prefixed, 24 in succession, to the chosen basis, with 
the element expressing the nearest relationship nearest the basis 
and that expressing the farthest relationship furthermost, until 
the desired relationship is reached. All sex indicators are suf- 
fixed. If the generation category of the structural basis is not 
apparent, as in chiu and i, the sex indicators also function as 
generation indicators; here, too, they are always suffixed. 

In choosing the structural basis for a term of a relationship, 
the factors to be considered are, first, generation and second, 
descent. Take, for example, the term for the father's father's 
sister's son's daughter's son. This is a complicated one, since the 

23 The sex indicators are fu %> mu #, tzu ^ , nil ^0 fu ^, fu #f , hsi #|, hsu 
iff. Failure to recognize this set of terms has resulted in much misunderstanding of 
the system. The first to disspell this misunderstanding was perhaps H. P. Wilkinson, 
Chinese Family Nomenclature, New China Rev. (1921) 159-191. He writes: "The 
initial error of the writers . . , was ... in taking the sex indicators for male and 
female appended to varying * descriptive ' appellations of kindred as the name of 
a class,— that of ' sons ' and ' daughters.' " A. L. Kroeber, quite independently, also 
discovered that "these last four terms (i.e. fu %, mu #, fu -^, fu #f) merely 
denote the sex of the person referred to, when they are added to other kinship terms. 
. . ." Process in the Chinese Kinship System, American Anthropologist 35 (1933) 151- 

* The terms " prefixing " and " suffixing " are employed here in a loose sense, since 
there are no true "prefixes" and "suffixes" in Chinese (with the exception, perhaps, 
of a few elements, notably the nominal suffixes). Here they merely indicate that a 
certain indivisible element (character) is placed before or after another indivisible 
element (character) in syntactic relationship. 


descent has shifted from female to male, and then back to female. 
Disregarding descent, let us first consider the generation. The 
individual concerned is of the son's generation. Instantly the 
basis is reduced to the alternatives chih or wai sheng. His imme- 
diate relationship with ego is through a female relative of ego's 
generation; therefore, the term chih is eliminated and only wai 
sheng remains. Furthermore, his relationship is a non-sib but 
consanguineal one, and descent is from father's father's sister, a 
relationship comparable with father's sister's ; therefore, the quali- 
fying elements ku and piao should be added. He belongs to the 
third collateral line of non-sib reatives; therefore, the collateral 
modifier fang is applicable. Together, these elements form the 
term t'ang ku piao wai sheng — a term as exact as can be desired. 
To express a female relationship of the same kind, add nil to the 
above term, making it t ( ang ku piao wai sheng nil [that is, f f si s 
d d]. To express a female relationship by marriage, substitute fu 
for nil; for a male relationship by marriage, substitute hsii. 

The elements which make up a compound term should always 
be interpreted in their extended, that is, their secondary, meanings, 
and should never be understood in their primary meanings. The 
amalgamation of all the extended meanings makes up the new 
connotation of the term so compounded. This phenomenon is a 
feature of Chinese syntax. Failure to understand this has been 
the source of much misinterpretation. 

The following illustrations represent practically the whole range 
of the structural bases. They are chosen with a view to including 
the widest variety of combinations, in order to elucidate the 
nature of terminological formations. The scope, however, is 
naturally limited, and fuller information must be sought in the 

Examples. The italics represent the nuclear term used as a 
structural basis, and the roman, the added modifiers: 

tsu if ku piao po f u f f si s > f 

po tsu mil ffobw s hu fyb 

t'ang shu tsu fu f f f b s < f f t'ang shu mu w of f f b *< f 

P° fob tsai ts'ung shu fu f ffbss<f 

fang po fu ffbs>f ku f si 


t'ang ku fu 
ku piao ku mu 
t'ang chiu fu 


m b 
mf bs 

t'ang tzu fu 
t'ang i piao tzu fu 
i mei 

h of f b d > e 
hof m f bdd>e 
y si 
wy si 

t'ang piao chiu fu 

m f f si s s 

t'ang i mei fu 


t'ang i fu 
tsai ts'ung i mu 

m si 

mf f bsd 

i piao mei fu 
t'ang i piao mei 


h of m si d < e 
m f b d d < e 



tsu chih nu 


ku piao hsiung 
t'ang ku piao hsiung 

f si s > e 
f f si s s > e 

tsai ts'ung chih hsii f f b s s d h 
wai sheng si s 


chiu piao sao 

t'ang chiu piao sao 

o b w 

w of m b s > e t'ang 

w of m f b s s > e 

t'ang wai sheng nil 
ku piao wai sheng 


f f si s d s 



chih sun nil 


t'ang ti fu 

w of f b s < e 

t'ang chih sun 

f bss s 

tsai ts'ung ti 

f f b ss<e 

wai sheng sun hsii 



o si 

In building terms for the third and fourth ascending and de- 
scending generations, the terms of the second ascending and 
descending generations are used as a basis, generation indicators 
are added to them. Modifiers of descent are usually added first, 
before the generation modifiers are prefixed. Examples: 

tsu f f sun s s 

tseng tsu mu f f m tseng sun nu s s d 

tseng po tsu fu f f f o b tseng chih sun fu b s s s w 

kao tsu fu f f f f hsiian sun s s s s 

The above represent the compositional principles of the 
standard system. 25 The standard terms are universal and form 
the patterns on which other terms are built or formed. They are 
for the most part used in formal — i. e., genealogical, legal and 
ceremonial — literature. In ordinary applications, they must be 
properly qualified by modifiers according to the specific situations 
under which they are used. 

Referential Modifiers 

The referential modifiers actually reflect the Chinese social code 
of etiquette, as well as the Chinese psychology concerning the 
proper attitudes to be assumed in social intercourse. It is a sign 

" Some would call it " literary system," in the broad sense of the term. 


of politeness and refinement to pay due respect and compliments 
to others, and, appropriately but not exaggeratedly, to maintain 
for oneself a more or less humble position. This is precisely the 
attitude that conditions the application of kinship terms. 

The referential modifiers are also a manifestation of the con- 
sciousness of membership in the relational group. The compli- 
mentary and depreciatory modifiers cannot be applied indiscrimi- 
nately; their application is prescribed by the identification with 
one relational group in contrast with another. Compliments may 
be applied more loosely, but depreciatives can be used only to 
those whom one strictly considers members of one's own relational 

These two attitudes are fundamental in the application and 
understanding of the whole terminology. 

The referential modifiers are governed by definite rules con- 
cerning their applications, and are always prefixed to the standard 
terms. With respect to their nature and usages, all of them can 
be broadly grouped under the following categories: 1. Compli- 
mentary, 2. Depreciatory, 3. Self -reference, 4. Posthumous. 

Complimentary. These elements are used in referring to the 
relatives of the person to whom one is speaking or writing. They 
consist of the following three elements: i. Ling fa: Illustrious, 
worthy, honorable. It may be prefixed to any standard term, 
except in instances where special stems are provided, ii. Tsun 
#: Honorable, venerable. Used synonymously with ling, but 
restricted in that it refers only to relatives of higher generation or 
status than that of the person to whom one is speaking, iii. Hsien 
Jt: Virtuous, worthy. Used alternatively with ling, but restricted 
in that it refers only to relatives of lower generation and status 
than that of the person to whom one is speaking. There are a few 
exceptions to this rule, e. g., hsien shu, " your virtuous paternal 

Whenever one is in doubt as to whether tsun or hsien should 
be prefixed, he uses ling. Ling, tsun and hsien have the sense of 
" your " used in a polite way. 

Complimentary modifiers should be prefixed when speaking to 
persons not related to oneself. They should not be used between 


sib relatives, except, when speaking to those of lower generations, 
in reference to their superiors. This latter practice is really 
teknonymy. The complimentary modifiers should be prefixed 
when reference is made to the relative of a non-sib relative to 
whom one is speaking, if that individual is not a connecting rela- 
tive. If he is a connecting relative and of higher generation than 
the speaker, the usual standard or vocative kinship term should 
be used. As a rule, one does not compliment those with whom 
one has close and direct relationships. 

Depreciatory. 2 * These modifiers are prefixed to the standard 
terms in referring to one's own relatives of the same sibname, 
when speaking or writing to others. " Depreciatory M is used here 
in the sense of " modest " or " of one's own." They consist of the 
following three elements: i. Chia %\ Family, dwelling, household. 
It is prefixed to the terms of all sib relatives of higher generation 
and status than ego. ii. She "&\ Cottage, shed, household. It is 
prefixed to the terms of sib relatives of the generation of, but of 
lower status than, ego (as younger brother) ; and principally in 
reference to sib relatives of the first descending generation, and 
sometimes all descending generations. It should never be used 
in reference to relatives in the direct lineal line, e. g., for one's own 
children, hi. Hsiao /!>; Minor, junior, small, diminutive. Prefixed 
to the terms of sib relatives of lower generation than that of ego, 
principally in reference to one's own children, grandchildren, etc. 
With the exception of the lineal descendants, she and hsiao can 
be used synonymously. 

Chia, she and hsiao have somewhat the sense of " my " used in 
a modest manner. It is important to note that depreciatory 
modifiers are not applicable to relatives of a different sibname. 27 
They are not even applicable to one's father's married sisters or 
one's own married sisters because these women have adopted their 
husband's sibnames and are no longer considered as members of 

"Depreciatory" is used in contrast to "Complimentary." As the elements chia 
and she show, " depreciatory " is used in the sense of " of my own family " or " of 
my own sib." 

,T There is a general term that can be applied to any non-sib relative, i.e., pi ch'in 
f$CfS, " my poor [or unworthy] relative." 


one's own family or sib, and therefore they are not to be 
" depreciated." 28 

Self -reference. 29 These modifiers are prefixed to the terms used 
by ego to refer to himself before another relative, either in speak- 
ing or in writing, e. g., a nephew refers to himself before an uncle, 
or vice versa. They consist of the following two elements: i. Yii 
M: Simple, rude, stupid. It can be prefixed to the terms when 
used by ego to refer to himself, principally as a relative of higher 
generation to one of lower generation, ii. Hsiao 'b: Junior, minor. 
It can be prefixed to the terms when used by ego to refer to him- 
self, principally where a relative of lower generation address one 
of higher generation. 

Neither yii nor hsiao are applicable to oneself where addressing 
a relative of the direct lineal line, e. g., father and son, grand- 
father and grandson, etc., where special terms are provided for 
such purposes. 

Posthumous. These modifiers are prefixed to — excepting a few 
special stems for this purpose — the standard terms when used in 
reference to one's own dead relatives, especially for parents, 
grandparents, father's brothers, etc. They consist of the follow- 
ing two elements: i. Wang t>. " Deceased." Prefixed to terms 
of all relatives when dead. ii. Hsien ^fc: " The late," " the former." 
Prefixed only to terms of relatives of higher generation or status 
than ego, when dead. 

When referring to the dead relatives of others the compli- 
mentary modifiers must again be prefix to these modifiers. This 
practice is not common; usually a circumlocutory expression is 

There are a number of special stems which are used with the 
referential modifiers. They will be pointed out in the tables (see pp. 
207-265) in each connection. For the sake of clarity and brevity, 
all terms qualified by the referential modifiers, or formed with 

28 Yen shih ckia ksiin jRKfg, 2- 5a /L#*»f , **, #, ^iPiP, B»M«* 

* Self-reference " modifiers are in a certain respect indistinguishable from " de- 
preciatory " except in context. It is especially true of the element hsiao. They are 
separated here for the purpose of exposition. 


special words, will be called in later discussion either compli- 
mentary, depreciatory, self -reference, or posthumous terminologies. 

Vocative Terms 

Vocatives are used as forms of addressing relatives direct in 
person. In literary address, i.e., in writing, the standard terms 
must be used. Vocatives must not be used together with referen- 
tial modifiers. The latter can only be prefixed to standard terms. 

Vocatives are limited to relatives of higher generations than 
ego, and to those of the same generation as, but of higher age 
status than, ego. Relatives of lower generations and age status 
can be addressed by name, or by using the standard terms as 
vocatives, if the occasion should arise. All vocatives are formed 
from three groups of terms: grandparent terms, parent terms, and 
older sibling terms. 

Grandparent terms. The grandparent vocatives vary a great 
deal with local usage. As they have not been systematically re- 
corded, it is rather difficult to determine the most prevalent ones. 
Yeh yeh, weng or weng weng, kung or hung kung, for paternal 
grandfather, p'o or p'o p ( o, nai nai, for paternal grandmother, may 
be considered the most common. No matter which terms are 
adopted in local usage, the adopted local terms are extended con- 
sistently throughout the whole system like these forms. In their 
extension, they are suffixed to the standard terminology by drop- 
ping the tsu fu and tsu mu, e. g., for po tsu fu (f f o b) the 
vocative is po weng, or po kung. 

Parent terms. Parent terms are less variable than grandparent 
terms. Tieh, yeh, and pa pa for father; ma and niang for mother. 
Pa pa is never, and niang is seldom, used in extensions. 

Tieh l£: Vocative for father. Used to form vocative terms for 
male relatives of the first ascending generation in place of fu. 

Ma iM: Vocative for mother. Used to form vocative terms for 
female relatives of the first ascending generation in place of mu. 30 

The above rules will not apply in instances where special 

30 Ma and mu, in their extensions, indicate a married status, and cannot be applied 
to unmarried female relatives. 


vocatives are provided. These terms may also be omitted in cer- 
tain cases where they are unnecessary, just as fu and mu are 
sometimes omitted. 

Older sibling terms. Ko, or ko ko -if If: Vocative for older 
brother. Used for conjugating vocative terms in place of hsiung 
for male relatives of the generation of, but older than, ego. 

Chieh, or chieh chieh JH.S&: Vocative for older sister. Used for 
conjugating vocative terms in place of tzu for female relatives 
of the generation of, but older than, ego. 

It is the vocative nomenclature that varies dialectically. At 
present, this variability mostly involves the grandparent and 
parent terms, the older sibling terms showing very little variation. 
But no matter how variable the dialectical vocatives may be, the 
above conjugation rules can be applied simply by replacing the 
given forms with local terms. 

The vocative terms are used more loosely, i.e., they are more 
" classificatory " than the standard terminology. When two rela- 
tives speak face to face the exact relationship is always under- 
stood ; it is only in referential usages that the more exact terms 
are needed. The prevalent use of sibnames, 31 personal names, 
titles, and numerical order of seniority and juniority 32 for par- 

sx Sibnames are used only for particularizing non-sib relatives and women married 
into the sib. 

82 The ancient method of denoting seniority and juniority by po f£], chung ^, shu 
)$£, and chi ^ has long been obsolete. A purely numerical order is used today. If 
ego's father is one of six siblings, A, b, C, D, e, and F (capitals indicate males, small 
letters, females) , the numerical order of ta ^C, erh H , san 2H » ssu , «m 3£ and 
liu 7^ will be applied to them, respectively. Ta is used in the sense of " eldest." 

Jih chih lu, 23.38a: ^A5t %f? * , IB— £:*, ^£o#}gTO. mm^M 

5Sffl-h^:5i , %'3Clft4f3¥%$—JiL . If ego's father is D, then ego will call A 
ta po, b erh ku t C san po, e wu ku, and F liu shu. If ego's father is A, then ego will 
call b erh ku, C san shu, D ssu shu, e wu ku, and F liu shu. The terms po and shu 
change positions in accordance with the relative order of ego's father, but the numerical 
order remains constant.* 

There is another method of assigning the numerical order, viz., by separating the 
male and the female series. As in the above case, A, C, D, and F will be assigned 
ta, erh, san and ssu, respectively, and b, e will be given ta, erh respectively. The 
method used depends upon local custom and family whim. 


ticularizing each relative in vocative address also makes the 
accurate system rather too cumbersome. 

Supernumerary Terms 

There are a few groups of terms which may be called " super- 
numerary," 3S viz., the sacrificial, epitaphic, literary and alterna- 
tive names. These are referred to in the tables on pp. 207-265. 

Sacrificial terms were used in ancient times for the direct lineal 
ancestors when offering sacrifices to them. There are only a few 
such terms, but they are now obsolete. Epitaphic terms are used 
on epitaphs and monuments. Strictly speaking, there are only 
two such terms, k'ao # for father and pi M for mother. It is 
only the sons who erect epitaphs for their parents. Sacrificial and 
epitaphic terms are often confused with terms modified by 
M posthumous " modifiers. They are frequently used interchange- 
ably, since they all refer to dead relatives, although in slightly 
different senses. Nevertheless, there are some very interesting 
changes which are of historical significance. 

Literary terms are those used only in literary compositions, 
usually non- vocative and non-referential. Many of them are old 
obsolete terms but still retained in literary usage. Alternative 
terms are those that can be used synonymously with the prevalent 
forms. The adoption of the one or the other depends entirely 
upon local custom and individual proclivities. 

u Supernumerary " is employed here in the sense used by E. W. Gifford in his 
discussion of California Kinship Terminologies, UC-PAAE 18, 1922-1926. It is not a 
happy term, and is adopted here only for want of a better one. 



The architectonic structure of the Chinese system is based upon 
two principles: lineal and collateral differentiation, and generation 
stratification. The former is a vertical, and the latter a horizontal, 
segmentation. Through the interlocking of these two principles, 
every relative is rigidly fixed in the structure of the whole system. 

Lineal and Collateral Differentiation 

The methods of differentiating collaterals differ in the ancient 
and the modern systems. In the ancient system, the Erh Ya and 
I Li, each collateral line is differentiated by following the termi- 
nology of the kin nearest to the lineal line from whom this line 
originated; e.g., father's father's father's brother is called tsu 
tseng wang fu, and his descendants down to ego's generations are 
differentiated by prefixing the term tsu to their respective terms ; 
father's father's brother is called tsung tsu wang fu, and all his 
descendants down to ego's generation are differentiated by pre- 
fixing the term ts'ung tsu. This method is also applied to more 
remote collateral lines. 1 

In the Erh Ya system there is no term for brother's sons and 
their descendants, nor is there any term for father's brother's son's 
sons and their descendants, nor for father's father's brother's 
son's son's sons and their descendants. It seems that the sons 
of brothers and sib-brothers merge into one another, i. e. brother's 
sons are one's own sons. On the other hand, the Erh Ya gives 
the term diu 2 for sister's son (man speaking) , li sun 3 for sister's 
son's son (man speaking) ; chih 4 for brother's son (woman speak- 
ing) , kuei sun 5 for brother's son's son (woman speaking) . In the 
strict patrilineal sib organization of the Chou period, even one's 
own sons are differentiated from one another as regards the order 

*Cf. Cheng Chen: ffilEWSfffl^fiJe. Ch'ao ching ch'ao "wen chi, 1. la-4b. 

9 ibid, mm-zvnmto- 

'Kid, ic?wm%±?®w, 

*ibid, 3BS£iP;8»«. 



of succession, hence it is difficult to see why there are no terms to 
differentiate one's own sons from brother's sons and sib-brother's 
sons, while, on the contrary, terms are provided whereby the man 
may differentiate his sons from his sister's sons, and the woman 
may differentiate her sons from her brother's sons. 6 

The differentiating of collaterals in the modern system is far 
more complete and consistent, but is carried out on a different 
principle. The generation stratum of ego is used as a basis, and 
the collateral modifying terminology is extended vertically down- 
ward and upward. E. g., father's brother's sons are called t'ang 
hsiung ti, their sons and grandsons are called t'ang chih and t'ang 
chih sun, respectively. Upwards, t'ang is extended to father's 
father's brother's son, e.g., t'ang po fu and t'ang shu fu; and to 
father's father's father's brother's sons, e. g., t'ang <po tsu fu and 
t'ang shu tsu fu. Other collateral lines, e.g., tsai-ts'ung and 
tsu, are similarly extended. 

The development of the modern principle of differentiation be- 
gan in the Han period. First came the differentiation of one's own 
sons from brother's sons by employing the terms yu tzu or ts'ung 
tzu. 7 During the Chin period the term chih was permanently 
changed from a woman's term for brother's son to a man's term 
for brother's son. Tung t'ang 8 was first used during the fifth 
and sixth centuries for denoting the second collateral line, and 
was later abbreviated to t'ang. Tsai ts'ung came into use a little 
later, and tsu is an old term used in a slightly delimited sense. 
With these important collateral modifying terminologies per- 
fected, the whole process was completed about the end of the 
first millennium A. D. 

8 It is very doubtful whether the Erk Ya system is complete. It also has no terms 
for fsiss, mbss, and m si s s. By inference, fsiss and mbss can be called ck'u, 
since sister's husband, f si s, and m b s are called sheng, and sister's son is called ch'u. 
Rut the absence of terms for m si s s is rather disconcerting; these terms cannot all be 
merged into the terms for ego's own sons, or into any others. For some reason or 
other the compilers of the Erk Ya seem not to have been interested in the terms for 
descendants of collaterals of the same generation. On the other hand, the Erk Ya 
system, as it stands, seems to stress the terms on the matrilineal side of descending 
generations. Whether or not this is a survival of an earlier matrilineate is a matter of 
interpretation, since other evidence is inconclusive. 

7 Cf. Table I, No. 125. 8 Cf. Table I. No. 41. 



Diagram I 

Ancient System of Collateral Differentiation 

f Kao tsu \ 

/ wang fu \ 
I Kao tsu J 

ywangrau / 

/Tseng tsu\ 
J wang fu ] 
I Tseng tsu J 
V wangmu J 

1 1 Tsu tseng \ 
' I wangfu J 

1 Wangfu ] 
I Wangmu J 

[ Ts'ung tsu \ 1 
I wang fu J I 

Tsu tsu \ 
wang f u J 

1 Mu ) 

[ Shihfu \ [ 
\ Shufu I \ 

Ts'ungtm \ 
fu J 

1 Tsutsufu ] 

( EGO J 

( K'un \ / 

V Ti / V 

Ts'ungfu \ 
k'un ti J 

rfef) (^k'untij 








Tseng sun 

Hsiian sun 

Italics indicate collateral modifiers. 


Diagram II 
Modern System of Collateral Differentiation 

Lineal 1st Collateral 2nd Collateral 3rd Collateral 4th Collateral 

Italics indicate collateral modifiers. 


Generation Stratification 

All relatives in the system are stratified in successive generation 
layers. This stratification performs the important function of 
fixing the exact location of relatives in the system, in conjunction 
with the principle of collateral differentiation. 

In diagram III (see p. 166) , the vertical columns represent 
collaterals and the horizontal columns represent generations. 
When one set of columns is superimposed upon the other, the two 
afford pigeonholes for every relative in the system. Each relative 
is then rigidly fixed, and not subject to fluctuations. The genera- 
tion strata are maintained by the use of generation modifiers. 
The modifiers are, in most cases, adapted from the nuclear terms 
from the direct lineal line, since in counting generations the lineal 
relatives are always used as absolute standards of measure. This 
adaptation of nuclear kinship terms as generation indicators has 
been interpreted as partial merging of collaterals and lineals. 
Given our present knowledge of the system, this interpretation is 
not tenable. 

Since generation is an important structural principle, it must 
not be disrupted, lest the structure break down. The most 
serious, if not the only, disruptive factor in this principle is inter- 
generation marriage. To counteract this influence, generation has 
become an important factor in the regulation of marital relations. 
A Chinese is not required to marry any of his or her relatives, but 
if such marriages do occur between relatives both parties must 
belong to the same generation stratum. In other words, a Chinese 
may marry any person outside of his or her own sib; if the 
parties are related, they must be of the same generation, irrespec- 
tive of age. 

This rule seems to have been less stringent in ancient times. 
During the Chou period a feudal lord could take his wife's paternal 
nieces as concubines, or even as wives after his wife's death. The 
Han emperor, Hui Ti (194-188 B. C.) , married his older sister's 
daughter, 9 and the Tang emperor, Chung-tsung (705-710 A. D.) , 

• Han shu frmm, 97A 5a: [#!«££] M«#ffiff8;&7C&±, 


married his paternal grandaunt's daughter. 10 These instances are 
severely condemned as incestuous by later historians and moral- 
ists, 11 but they were not so condemned by contemporaries. On 
the other hand, these instances may have been anomalous; it 
may be only because the marriages involved emperors that they 
went unpunished and uncriticized. But, in either case, they do 
show the laxity of the generation rule during the earlier period. 12 
There can be no doubt that the generation rule was much 
stressed even during the Chou period, 13 since the recorded marri- 
ages show that the inter-generation type of marriage was the 
exception rather than the rule. 14 Its stiffening was gradual, and 
culminated about the middle of the first millennium A. D. The 
period of intensive development of the principle seems to have 
been about the third and fourth centuries A.D., because it is 
during this period that the generation indicators in personal names 
became popular. The Tang Code (ca. 600 A.D.) contains 
clauses which definitely prohibit marriage between relatives of 
different generations. 15 All subsequent codes contain such inter- 
dictions, and cite cases. From the end of the first millennium A.D. 

rTT^3R:±, $$JnMH 9 

"Wang Ming-sheng (1723-1797 A.D.) discussed these instances in his-f-^ jj»gj 
fiKjUffl^Bf^i^O 86.2a: as the most flagrant violations of the "relationships of 
humanity," AflfctSIB. 

13 The Tung tien discussed two instances of difficulties in mourning obligations 
arising from inter-generation marriages, 95. 12a: jJSXIbSl %, MWiWk J and the hypo- 
thetical case, 95. 5b-6b: ^M^f^^iicM^ y jftf tlti&Wk . It seems that inter- 
generation marriage between distant relatives was tolerated during the first half of the 
first millennium A. D. 

13 Both the tsung fa and the sang fu institutions, which were developed during this 
period, stress the separation of generations. Sang fu will be dealt with later. As to 
tsung fa, the subdivisions of tsung into minor tsung /J\ ^ , for the most part, depends 
upon- the counting of generations. 

14 The instances of primary inter-generation marriages can be counted on the fingers. 
Liang Yii-sheng [P'ieh chi, 2. 2a] says: ^$UR£¥~£, (ifeflUr— ) , §£ 

&mm& as-if-ra), ««»«##, -smmm*. mmm^, «# 

"J'anglushul J3Jf, 14. 2a: j£ft*HWflBJH , ffif#**«*l«i • • • JK 
38c ii&. This is followed by expositions of this clause, and by another clause of the 
same nature but more specific. H=£Q means " relatives belonging to different genera- 



Diagram III 

Descended from females through males 

Descended from males through males 

The heavy squares represent the nuclear group of relatives. Those in italics, indicate their descendants have 
not been carried over into the next generation, e. g., the children of nu are wax sun and wai sun nil but not given 
in the following square. The Roman numerals represent ascending and descending generations. 


to the present, not only have inter-generation marriages been 
rigorously forbidden by law, but popular sentiment against them 
runs so high that even a teacher marrying his or her pupil, or a 
person marrying a friend's daughter or son, is condemned. 

The underlying concept is the desire to keep constant the 
generation layers of relatives and to prevent their disruption. If 
the generation of one relative is disrupted by marriage, then all 
the positions of the relatives connected with him would also be 
disrupted, and the system would lose its accuracy of description 
and thus defeat its own purpose. 


Kroeber's essay on the Classificatory Systems of Relationship lfl 
does not invalidate the use of classificatory and descriptive desig- 
nations in anthropological discussion; its contribution lies in the 
establishment of categories that are inherent in all systems. These 
categories constitute a convenient means for examining the work- 
ing processes of any system. It is impossible to tabulate the 
whole Chinese system into one table, but we may take the nuclear 
terms and tabulate them in the light of the eight categories 
suggested by Kroeber. It must be understood that these nuclear 
terms are also used in secondary meanings, qualified by modifiers. 
In this tabulation (see p. 168) only their primary meanings are 

Now let us consider the system as a whole, together with the 
nuclear terms under each category. 

lfl Journal, Royal Anthropological Institute of Great Britain and Ireland 39 (1909) . 



Percentage | | 1 | 

i> i> © O «5 GO 

Tol'll 00 00 CO CO CO © GO 

$£g f U * * * * * * 

* * * 

w "3 &P 




8? sao ***** * | 

ft! ch'i * ? 

^K fu * ^ 

3£ hsii ***** g 


1\J= y Q ***** W 

Jg J * * * ^ 

S chiu * * * o 

£ ku * * * 

E sheng * * 

!g chih ***** o 


& shu * * * * < 

9 po * * * * I 

3S mei * * * * * „ 

@ tzu * * * * * % 

$£ ti **** * ^ 


S? hsiung * * * * * § 

•^ nii * * * * ^ 

^ mu * * * * gj 

tf* tzu * * * * | 

« f u * * * * 2 

m sun ***** 3 

fig tsu ***** p* 

I ! I 1 . J 8 1 1 

i i 1 1 Mi 

.Si -2 l. fe 13 ° pS 9* & o &c 

SP fc "S 1* ^ ° ^ -a -a -g 

+2 3 X fi * u . i) fi « 

Is § .S .3 8 S S Sf r <? a 

tj O M h! tfi tf) y2<J^ 


i. The difference between persons of the same and of separate generations. This 
category is rigorously observed in the whole system through the use of generation 
indicators. Generation is not only an important structural principle in the system but 
is also an important regulator of marriage and a determinator in the application of 
vocatives. But in the nuclear terminology it is represented by only 78.3 per cent, of 
the terms. In the terms po, shu, ku, chiu, i the generation category is overridden. 
This merging of generations is not inherent in the system, but has been produced 
through the disruptive force of teknonymy, which will be discussed later. 

ii. The difference between lineal and collateral relationships. This category is strictly 
fixed in the whole system. Collateral lines are differentiated by special modifiers. 
Practically all the basic modifiers exist solely for the function of developing this 
category. In the nuclear nomenclature it is represented by 100 per cent of the terms, 
iii. The difference of age within one generation. This category is only partially 
represented in the system. It is fully represented in ego's own generation, whether 
male or female. Among the ascending generations it is operative only among the male 
relatives, with the exception of the wives of male relatives. Among the descending 
generations it is not operative at all. In the nuclear terminology it is represented by 
only 34.8 per cent. 

W. H. R. Rivers attributed the differentiation of age within one generation to the 
practice of tribal initiation, i.e., older brothers will be initiated before younger 
brothers. 17 Whether or not this be correct, it is the only general explanation seriously 
advanced. In ancient China there are vestiges of initiation rites, especially as 
recorded in the J Li ls and Li Chi. 19 Whether or not these ancient initiation rites 
have anything to do with the expression of this category is by no means certain. 
Chinese authors usually connect it with the sib organization, tsung fa, since in this 
organization the older brothers have absolute priority over the younger brothers in 
the transmission of office and property, and special privileges in the sacrificial rites 
to ancestors and in carrying on the line in general. 20 

iv. The sex of the relative. This category is consistently carried out in the whole 
system through the employment of sex indicators. In the nuclear terminology the 
representation is 100 per cent. 

v. The sex of the speaker. This category is entirely inoperative. The sex of the 
speaker is always understood but never expressed. Rut there are traces of this category 
in the ancient system. E. g„ in the Erh Ya and the / Li, the term chih is exclusively 
used as a woman's term for brother's children. A few other terms may have been used 
only as women's or men's terms, but here we are less certain. 

vi. The sex of the person through whom the relationship exists. This category is 
fully expressed in the system through the use of special modifiers. E. g., piao hsiung ti 
may mean father's sister's sons, mother's brother's sons, or mother's sister's sons. But 
if we say ku piao hsiung ti, chiu piao hsiung ti and i piao hsiung ti, the terms are 

1T W. H. R. Rivers, Socio 1 Organization, ed. by W. J. Perry, 1924, p. 189. 

19 mm- _ 

^Ch'eng Yao-t'ien ^SgBB (1725-1814): Tsung fa hsiao chi ^^/>g£ 1.1a: 


exact, and refer to father's sister's sons, mother's brother's sons and mother's sister's 
sons, respectively. There are lapses in the vocative usages, since the exact status of 
the connecting relative is always understood and never expressed. 

vii. The distinction of blood relatives from connection by marriage. With the 
exception of the terms chin and t, this category is consistently expressed. The merg- 
ing of mother's brother (consanguineal) and wife's brother (affinal) into chin, and 
of mother's sister (consanguineal) and wife's sister (affineal) into i, is due to the 
influence of teknonymy. In the vocative usages the terminology is looser, because 
of the mutual adoption of each other's terms by husband and wife. 

viii. The condition of life of the person through whom relationship exists. This 
category is present, but usually is not consistently expressed. The most common dis- 
tinction is between the dead and living relatives, by the application of special modi- 
fiers. There are a few distinct terms for this purpose for parents, grandparents, 
paternal uncles, etc. Other conditions of life of the relative are not expressed, or are 
indicated only by circumlocutory expressions. 

Certain of these categories (ii, iv, vi and possibly vii) are essential to the maintenance 
of a strict unilateral — patrilineal in Chinese — descent. It is also exactly these categories 
that find their fullest expression in the Chinese system. The great nicety in the 
distinction of dead and living ancestors (category viii) in the ancient terminology may 
have been due to ancestor worship, which is less prevalent now. 


In kinship systems there is usually the factor of reciprocity to 
influence the expression of certain categories. Reciprocity is of 
three kinds: logical or conceptual reciprocity, verbal reciprocity, 
and self-reciprocity, i. e., both conceptual and verbal. 21 It is not, 
on the whole, a feature of the modern Chinese system, since, to a 
certain extent, it is incompatible with the consistent expression 
of certain categories and exactitude in the discrimination of 

In the ancient system there are traces of conceptual reciprocity. 
The / Li 22 says, " They call me ku, I call them chih" Inversely 
it is also true: "They call me chih, I call them ku!' Ku, as 
used in the Erh Ya and the J Li, means father's sister, and chih 
means brother's child (woman speaking) , both male and female. 
In other words, ku " indicates the sex of the relative but not of 
the speaker," whereas chih " does not recognize the sex of the 

21 Cf. the definition of reciprocity suggested by A. L. Kroeber, California Kinship 
Systems, UC: PAAE 12, 9.340, note 1; Zuni Kin and Clan, Anthropological Papers, 
AMNH 18 (1919). 78-81. 

2a mm «■ ib: ®m : «*, fir*? tnmm, ^m^m. 


relative indicated, but does imply the sex of the speaker." 23 
Therefore, ku and chih each involve a category which the other 
does not express. Ku and chih in their ancient usages are true 
conceptual reciprocals. 

The Erh Ya also states, " They call me chiu, I call them 
sheng" 24 Chiu is here used in the sense of mother's brother, and 
sheng, sister's child. But the reverse does not hold true. Actually 
this statement contradicts the other terms recorded in the Erh 
Ya itself, since sheng was used in the sense of father's sister's son, 
mother's brother's son, sister's husband (man speaking) , and 
wife's brother. The Erh Ya also gives another term, chu, for the 
same relative, sister's child. It is very likely that the above state- 
ment is a later interpolation, since chiu-sheng was probably re- 
ciprocal from the Han to the T'ang periods. 25 The whole problem 
is complicated by the question of cross-cousin marriage and 
teknonymy, which will be dealt with later. In short, chiu-sheng 
could not have been reciprocal in the ancient system, but sheng 
was itself partially reciprocal, i. e., in the male sex only. 26 

These are the only traces of reciprocity that can be detected in 
the ancient system. The tendency toward the consistent use of 
categories later became so strong that even these few vestiges of 
ancient reciprocal terms have entirely disappeared from the 
modern system. 

23 A. L. Kroeber: Classificatory System of Relationship 81. 

24 This statement also occurs in the I Li, ^M 33.9a: j|0: i%$£ > fR^? fl 

25 During this period chiu was used for mother's brother alone, and sheng dropped 
all its other connotations and became simply a term for sister's child, e. g., wai-sheng. 

28 That is, sheng is used reciprocally between mother's brother's son, father's sister's 
son, sister's husband (M.S.), wife's brother, and male ego. They call male ego sheng, 
and male ego also calls them sheng. 


In the course of development from the ancient to the modern 
system there has been a slow but persistent tendency toward 
systematization and more exactly descriptive efficacy in nomen- 
clature. This tendency has to a large extent been conditioned by 
the sociological lines along which Chinese society is organized, 
and has to a lesser extent been intentionally fostered by ardent 
ritualists and framers of etiquette. On the other hand, there are 
also potent social forces which work against overrationalization 
and oftentimes throw certain parts of the system out of gear by 
undermining certain categories. The systematizing forces are the 
sib-organization and the mourning system. These two factors 
have supplemented one another in moulding the kinship system to 
their own pattern. The disrupting forces are the Chinese marriage 
customs and, most potent of all, teknonymy. The ritualists 
served as a stabilizing agency in vigilantly conserving the nomen- 
clature, 1 with a view to the exactitude of mourning specifications 
and the needs for maintaining sib solidarity. But generally they 
were powerless against the popular tendencies in kinship usages, 
and very often were forced to accept the already established 
terms and attempt to harmonize and incorporate them into the 
whole system. 2 Thus, in the Chinese system there is a seeming 
embodiment of a well considered plan while, at the same time, 
there are many incongruities. No matter how much we may dis- 
credit " mock kinship algebra," to borrow a phrase from Mali- 
nowski, 3 it is pertinent to inquire into the conflicting forces which 
have moulded, and are today shaping, the Chinese system. 

1 For example, chiu and ku as terms for the husband's parents have been obsolete 
since the turn of the first millennium A. D„ but they are still used in this sense in 
ritual works. Chiu was extended to include wife's brother ca. 900 A. D., but one 
never finds chiu used in this sense in formal literature. 

2 E. g., i was extended to include mother's sister ca. 500 B. C. During the next seven 
or eight centuries, the new and old terms were used interchangeably. About 400 A. D. 
t was standardized as a term for both mother's sister and wife's sister. 

3 Kinship, Man 30 (1030) . 17. 19-29. 



The Sib: Descent and Exogamy 

Kinship ties necessarily begin within the family as a pro- 
creational unit. These primary ties, as we may call them, 4 are 
biologically the same in all societies, 5 though, functionally, they 
may differ from culture to culture. 6 But kinship ties do not rest 
within the reproductive family. They are extended to a much 
wider circle of individuals who are actually or reputedly related 
to those of the procreational family. In this process of extension 
certain groups of related individuals are emphasized and certain 
others minimized, although their degree of relationship may be 
exactly the same. The basis of this variability in the grouping 
of kin is the subject of much, perhaps unduly much, anthropo- 
logical discussion. And naturally so, for it is precisely this vari- 
ability in kinship patterning that differs so widely among differ- 
ent peoples — particular systems grouping relatives in quite differ- 
ent ways. The character of the kin groups emphasized likewise 
reflects the wider ranges of the social structure of which the kin- 
ship system is part. 

In the Chinese kinship system, relatives in the male line receive 
emphasis ; the formalized basis of which is the exogamous patri- 
lineal sib. 

Sib organization is called tsung fa in Chinese, literally, the 
" law of kindred." The tsung fa was bound up with the feudal 
system, 7 which was swept away in the course of the third century 
B. C. The sib organization, however, has survived to the present 
day, although in a much attenuated and modified form. The post- 
feudal development of the sib reached its climax in the third to 
the eighth centuries A. D., when it is usually termed shih tsu or 

* B. Malinowski would call this " the initial situation of kinship ": preface to 
Raymond Firth's We, The Tikopia [1936] x. 

8 I. e., the divisions into father, mother, son, daughter, brother and sister, members 
of the strictly procreational family, exist in all kinship systems. 

8 E. g., the relationship of " father " to the rest of the procreational group in a 
patrilineal society may be radically different from that in a matrilineal society. 

7 For the sib organization of the feudal period, cf. Wan Kuang-t'ai fS^^t^, Tsung 
fa -piao xjt^feiSt; Wan Ssu-ta ^^f^C, Tsung fa lun T^^feffe; and Ch'eng Yao-t'ien 
f=|J§&EH» Tsung fa hsiao chi T^^fc/J^pfi ■ For a more modern study, cf. Sun Yao 

MM> Ch<un ch ' iu shik to * chih shih tsu ^%k$fft^WM [1931]. 


tsung tsu. The causes of this excessive development were many, 
but primarily it represents a reactionary growth following the 
abolition of the feudal system. The larger and more prominent 
sibs took the place of the feudal nobility, both in monopolizing 
governmental offices and in maintaining social prestige. 8 Their 
influence began to decline during the T'ang period. This was, on 
the one hand, due to the suppressive measures of the T'ang rulers 
and, on the other hand, to the social upheavals precipitated by the 
decline of the T'ang dynasty. 9 At the present time sib organiza- 
tion, for most people, is less vital than formerly, but its tradi- 
tions and influences still permeate the whole of Chinese social life. 
Tsung fa itself has been well studied, and we need here con- 
sider only those two of its characteristics which have direct in- 
fluence upon the alignment of relatives, viz., patrilineal descent 
and exogamy. With the tsung, the sib, each line of descent is 
strictly patrilineal, not only in the transmission of the sibname 
but also in the transmission of office, property, etc. It is also 
primogenitary: the eldest brother has priority over the younger 
brothers. 10 The tsung is absolutely exogamous. Marriage within 
it is impossible, even after a " hundred generations." It is also 
strictly patrilocal. 11 Evidence is far from conclusive as to whether 
or not Chinese society passed through a prior matrilineal phase. 12 
Exogamy, however, was predominately a Chou institution, 13 and 

8 Cf. Kai yii ts'ung ttao, ^fBlf/Ht^, 17. lb-9a. 
B Cf. Tung Chih jg-g : Rjj£B£#, 25.1. 

10 In the transmission of hereditary titles, the primogenitary principle holds but the 
property is divided equally among the brothers, although the eldest brother usually 
receives an extra share. It is only in the ceremonies in the ancestral halls that the 
primogenitary line has absolute priority over all collateral lines. 

11 When, for want of male issue, a son-in-law is adopted as son, he adopts the 
wife's sibname and lives with her parents. He is in every way treated as a son. The 
children of the third generation revert to his original sibname .nf^IU^, but usually 
one male child is allowed to carry on the wife's line. 

12 The existence of an early matrilineal state constitutes the underlying hypothesis 
of many recent works on ancient Chinese society, e. g., M. Granet's La civilisation 
chinohe, Paris, 1929, and Kuo Mo-jo's t t , ®"j&^fflh'^'^F!S» 1931. The evidence 
they have marshalled is suggestive rather than conclusive. 

»u chi *®, 34.7b: %~z)>mmn, mziatrnftft, ars"M# 

*@^M*, ft £&«&!!• . T'ung Tien, 96.2a, cites WIESE^A: JLHIJffi'E'ft 


its development was rather late. Ancient authorities attest that 
during the Hsia and Shang [ca. 1700-1100 B. C] periods members 
of the same sib could marry after the lapse of five generations. 14 
The institution of strict sib exogamy was traditionally attributed 
to Chou Kung [ca. 1100 B. C], who instituted it for the main- 
tenance of sib solidarity. Nevertheless, there is abundant evi- 
dence to show that even during the Chou period this interdiction 
was neither universal nor strictly enforced. 15 It was only after 
the overthrow of the feudal system and the transformation of the 
sib organization that absolute sib exogamy gradually prevailed. 
From the middle of the first millennium A. D. to the present this 
rule has been vigorously enforced by law. 16 

The moulding effect of the exogamous patrilineal sib is seen in 
the dichotomy of relatives in the system. Relatives are divided, 
along sib lines, into sib relatives [tsung ch'in] and non-sib relatives 
[wai ch'in or nei ch'in]. All sib relatives belong to the same sib 
of ego and possess the same sibname. Paternal relatives descended 
from females through males or females are non-sib relatives. 
Maternal and affinal relatives all belong to the non-sib group. 

To maintain this distinction between the sib and non-sib groups, 
the terminology must be bifurcated in such a way that the 
paternal relatives descended from males through males are differ- 
entiated from those descended from females through males. This 
is carried out by differentiating the father's brother's sons [t'ang 
hsiung ti] and their descendants from father's sister's sons [piao 
hsiung ti] and their descendants ; brother's sons [chih] from sister's 
sons [wai sheng] ; son's sons [sun] from daughter's sons [wai sun] ; 
and so on, ascendingly, descendingly, and collaterally. This bi- 
furcation is necessary, since father's sister, sister, daughter, or any 
female sib relative who through sib exogamy has married into 

14 Tai p'ing yii Ian Jk'P'ffl'M [1808] Hft ed., 540. 7b-8a, cites jjjft^MS. The 
five generations include the generation of ego. Even then, it is still very doubtful whether 
the Hsia and Yin peoples had any exogamy at all. 

15 Chao I says [Kai yu ts'ung k'ao, 31.2b-3a], IpJ^&^f , Miti^%k$?M%> and 
cites many cases to support his thesis. He concludes, itfc^^^Cff^^LI&^iL , §!<$ 

= "E.g., in T'ang lu shu i 14.1a: ftf^tf#j&£, frf&ZL*?, 1&M.\&±\>i1& 
jifo. Cf. Pierre Hoang: he manage chinois au point de vue legal, [18981 43-53. 


other sibs, and their descendants, belong, on account of patrilineal 
descent, to different sibs from ego. 

Nevertheless, the emphasis upon the sib relatives is not so 
manifest in the whole kinship system as it might be. A partial 
explanation lies in the minute differentiation in the terminology, 
which to a certain extent has obscured this grouping. If we look 
at the application of the depreciatory modifiers, however, this 
emphasis at once becomes apparent. Only sib relatives are to be 
depreciated, inasmuch as they are regarded as members of one's 
own group. Non-sib relatives are not to be depreciated, since 
they are felt to be outside of one's own group. 

Conceptually, sib relatives are considered nearer than non-sib 
relatives, even though their degree of remoteness from ego may 
be exactly the same. This is best expressed in the mourning 
obligations. The mourning period for paternal grandparents is 
one year, but for maternal grandparents only five months; for 
paternal uncle one year, and for maternal uncle five months ; for 
father's brother's son nine months, and for mother's brother's son 
only three months. More instructive is the difference in mourning 
periods for a female sib relative before and after her marriage. 
The mourning period for father's unmarried sister, ego's unmar- 
ried sister, unmarried daughter, and brother's unmarried daughter, 
is one year ; after their marriage, the mourning period is decreased 
to nine months, i. e., it is lessened by one degree. Therefore, we 
see that as long as these females remain unmarried they belong 
to ego's sib, but after marriage they belong to their husband's sibs. 
This transference through marriage has lessened their bond with 
the sib and, consequently, with ego. 17 

The nomenclature of the non-sib relatives is perhaps more 
expressive of this dual division of relatives. The differentiation 
between paternal grandparents [tsu] and maternal grandparents 
[wai tsu], paternal uncle [po and shu] and maternal uncle [chiu], 
paternal aunt [ku] and maternal aunt [i], is, of course, a regular 

17 The mourning obligations of an unmarried female to her paternal relatives are 
just the same as those of her brother. After marriage, all these obligations are lessened 
by one degree, and, reciprocally, obligations of these relatives to her are also lessened 
by one degree. 


feature of a system based upon exogamous social grouping. But 
the interesting phenomenon is the merging of the father's sister's, 
mother's sister's, and mother's brother's descendants in the single 
term, piao. 1 * Piao, as a term in itself, means " outside " or 
" external." The descendants of father's sister, and of mother's 
sister and brother, although consanguineal relatives of distinct 
affiliation, are all non-sib relatives, and hence their merging in 
the term piao is understandable. 

The introduction and development of piao is also of historical 
interest. In the ancient system of the Erh Ya and the / Li, 
father's sister's sons and mother's brother's sons were merged in 
the term sheng, through cross-cousin marriage. 19 Mother's sister's 
children, being parallel cousins, stood alone as tsung mu hsiung ti 
[for males] and ts'ung mu tzu mei [for females]. During the first 
two centuries A. D., with cross-cousin marriage already long in 
abeyance, father's sister's sons were designated as wai [outside], 
e. g., wai hsiung ti 9 20 and mother's brother's sons as nei [inside], 
e. g., nei hsiung ti. 21 At about the same time 22 chung and piao 
were used as equivalents of nei and wai, since chung means 
" middle," " inside," and piao means " outside," " external." 

During the last few centuries of the first millennium B. C. and 
the first few centuries A. D., mother's sister's sons were usually 
designated by the newly extended term i, e.g., i hsiung ti, al- 
though ts'ung mu hsiung ti was still permissible. 

From the fourth to the seventh centuries A. D. constant con- 
fusion was produced through the use of the terms nei or wai for 
mother's sister's children, and, consequently, a confusion of rela- 
tives involving father's sister's, mother's brother's and mother's 
sister's descendants. 23 The reason for this may lie in the fact 

18 The ku piao, chin piao and i piao are the three first-degree piao relationships of 
the Chinese system. 

19 See discussion under Cross-cousin Marriage, below, pp. 183-186. 
90 I Li M%> 33.9b: Jfrfc^. »&: &H.&*.. 
«/Wd.,SS.10b: Jt^. #&: ft5&#-fe. 
n H<m-Hanshu%±i&,100.*ar$*: • • • ^Hfll, 7516^1 M»H •" • • # 

83 E.g., Tuns rien,95.9a-12b: Sft&ttSJtf&lfJJftW: (The instance involved 
is neither mother's brother's daughter, nor father's sister's daughter, who ought to be 


that these relatives, although of clear affiliation, all belong to the 
non-sib group and are distinct from the exogamous partilineal sib 
group. During the Tang period, the chung was dropped and 
piao alone was applied to all these relatives. 24 Thus piao became 
a general indicator for non-sib consanguineal relations descended 
from relatives of higher generations than ego. 

This conceptual as well as nomenclative dichotomy of relatives 
is a definitive expression of the exogamous patrilineal sib principle. 
At the same time, this principle has been modified by another 
factor — mourning rituals — resulting in the elaborate differentia- 
tion of collateral lines in the sib group, rather undermining its 
original function. But an understanding of this principle is most 
essential for a grasp of the system as a whole, since it is not only 
a potent moulding force but is as well a controlling factor in many 
important kinship usages, i. e., the depreciatory and complimentary 

Mourning Grades 

The Chinese mourning system is based upon the sib organiza- 
tion for its discrimination against non-sib relatives, and on degree 
of relationships for the assignment of mourning grades. Mourn- 
ing for sib relatives vanishes at the fourth degree [fourth collateral] 
and at the fourth generation, both ascending and descending from 
ego. 25 There is, consequently, in the kinship system a sharp 
differentiation of the first four collateral lines, and an indefinite 
grouping of all further collaterals in the tsu relationship. 

Mourning, sang ju in Chinese, is a colossal subject in itself; 

nei or wai, respectively, but mother's sister's daughter) ^W?$aW?Z : ^MnlSc 

M$%W^ (cited by CV&iflieetfu, 3.20a) states: MAMtM'Z.^, ffitM1$fl*!rl 
fc. And lI]A$#, ftm, *• ?b: ffi&£ : f#ftJZ.&» & Jl^^ft ft 
&. -K, 0?»#*;»&5aS&.*6?»M : P;Bftfl&. Theuseofnei 
and wai is so confusing that even these encyclopaedists do not know which usage 
is correct. [The missing character is No. 16 on page 149.] 

24 During the sixth century the subject of piao relationships became so popular 
that even genealogies were compiled for them, e.g., Wei shu rftlS13» 57.5a: ig^ 

» Li Chi it® , 34. 7b: EiftMlg , JiB^SMl . 


here we can only touch upon those fundamental aspects that are 
prerequisite to an elucidation of its influence upon the kinship 
system. Chan tsui (three years) , tzu tsui (one year) , ta kung 
(nine months) , hsiao kung (five months) , and ssu ma (three 
months) are commonly known as the wu ju, " five grades of 
mourning." Actually, the number is greater than five. The 
specifications of these grades have fluctuated much from. period 
to period; certain grades have been dropped or added in con- 
formity with the excentricities of particular periods. Although 
the specifications may thus have changed, the fundamental prin- 
ciples which underlie these specifications have remained constant. 
All the paraphernalia expressed by the above terms 26 are mere 
accessories, the fundamental units of mourning being the " mourn- 
ing periods." The basic unit is the year. All other degrees are 
either chia lung [increased mourning], or chiang shai [decreased 
mourning], relative to the basic unit. These principles are best ex- 
pounded in the San nien wen of the Li Chi, 27 which says: " Why 
is it that the mourning period for the nearest kin is one year? Be- 
cause the interaction of heaven and earth has run its round ; and 
the four seasons have gone through their changes. All things be- 
tween heaven and earth begin their processes anew. The rules of 
mourning are intended to resemble them." " Why should there 
be three years mourning [for parents]? The reason is to make it 
more impressive, chia lung, by doubling the period, so that it 
embraces two round years." 28 " Then why have the mourning 
of nine months? The reason is to prevent excessive grief." There- 
fore three years [actually twenty-five months counted as three 
years] is the highest expression, lung, of mourning. Three months 
and five months 29 are the lowest, shai. One year and nine months 

28 The terms chart ts'ui, tzu ts'ui, ta kung, hsiao kung, and ssu ma specify the kind 
of apparel to be worn at the mourning for a particular relative. 

27 ---^ppjj^ 58. 3a-4b. For translations in European languages, cf. J. Legge, Li Ki 
Sacred Books of the East 28.393-4; and S. Couvreur, Li Ki [1913], 580-586. 

28 Twenty-five months are counted as three years, hence the " three years " mourn- 
ing is only two years and one month. 

28 " Five months " are counted as two seasons, which should be six months. The 
substitution of five for six reflects an ancient Chinese aversion for even numbers. Thus, 
the basic periods are: three months (one season), five months (two season), nine 
months (three seasons) , and one year (four seasons) . 


are the norms. Heaven above gives an example ; earth below, a 
law; and man between, a pattern. The harmony and unity that 
should characterize men living in their kinships are hereby com- 
pletely shown." 

Mourning starts from the nearest kin with the basic unit of 
ch'i, one year. The nearest kin have three. According to the Sang 
fu chuan,™ " the relation between father and son is one ; between 
husband and wife is another; and between brothers is a third." 
With the three nearest kin, with the basic unit of clxi, and with 
the principles of chia lung and chiang shai, the whole system is 
correlated with the kinship system as in Diagram IV (see p. 182) . 

The process is as follows: The mourning period for father is 
ch'i, one year ; for grandfather, ta kung, nine months ; for great 
grandfather, hsiao hung, five months ; and for great great grand- 
father, ssu, three months. This is called shang shai, " ascending 
decrease." The mourning period for son is ch'i, one year; for 
grandson, ta hung, nine months ; for great grandson, hsiao kung, 
five months; and for great great grandson, ssu, three months. 
This is called hsia shai, " descending decrease." The mourning 
period for brother is ch'i, one year ; for father's brother's son, ta 
kung, nine months ; for father's father's brother's son's son, hsiao 
kung, five month ; and for father's father's father's brother's son's 
son's son, ssu, three months. This is called pang shai, " horizontal 

The chan ts'ui, three years, mourning for father is chia lung, 
increased mourning; basically it is only one year. The one year, 
ch'i, mourning for father's father and father's brothers, is likewise 
chia lung, since basically it is the ta kung grade, nine months. 

All non-sib relatives, whether consanguineal or affinal, are given 
the last grade of mourning, ssu ma, three months, no matter how 
closely related they may be. 31 They are subject also to the prin- 
ciple of chia lung-, e.g., in the ancient mourning specifications, 
the mourning grade for mother's sister is hsiao kung, five months, 
but for the mother's brother it is only ssu ma, three months ; 32 

80 / Li 30. 8b-9a. 

81 / U ftjffi, S3- 2a: ^MH^DR-g&Ufc. Cf - Jih chih lu 5S5 ' 7 ' 

83 In modern mourning specifications mother's brother is increased to five months, 
hsiao kung. 


the former is chia lung, and the latter is not. Basically they are 
all ssu ma, 33 

Mourning grades of a simpler kind must have existed long be- 
fore the Chou period, but their elaboration began only when they 
fell into the hands of the Confucianists. 34 Using the family and 
sib as the bases for their ideological structure, these literati elabo- 
rated the mourning system with a view to the maintenance of sib 
solidarity. In the course of this elaboration of the mourning 
system they also standardized its basis, the kinship system, for 
a carefully graded system of mourning rites requires a highly 
differential kinship nomenclature, lest an awkward incommen- 
surability ensue. This is especially apparent in the comparison of 
the Erh Ya and the / Li systems. The Erh Ya system, when 
compared with the system recorded together with the mourning 
rites [Sang ju chuan] in the / Li, is inconsistent and less differential 
in many respects. Certain classical scholars naively tried to amend 
the Erh Ya with the / Li system, since they considered the Erh 
Ya system below the standard of Confucian ideals of kinship. 35 
They failed to see that the Erh Ya represents an early state of 
the system, and the I Li & later but rationalized system worked 
over to conform with the mourning system. 

There is no doubt that the Erh Ya system was already to some 
extent rationalized through Confucian influences, but it is much 
less so than that in the I Li. With the Confucian ideals firmly 
implanted in the Chinese social structure from the second cen- 
tury B.C. on, the mourning rites were increasingly elaborated 
and popularized, 36 and concomitantly the kinship system, until 
both reached their apogee during the T'ang period. 

The elaborate mourning rites are a distinctive feature of Chinese 

33 For the actual specifications one must consult the ritual works of each period. 
The above are merely general statements of principle. 

84 Certain scholars believe that the three-year mourning for father was a Shang 
custom, and that the Ju §j, who practiced it during the Chou period and later 
evolved into Confucianists, were descended from the Shang dynasty people. Cf. Hu 
Shih: gftfit, $%$i^^7!i^t 9 j$— My 19-23, 90-94. For an opposite view cf. 
Feng Yu-lan, The Origin of Ju and Mo, CHHP 10.279-310. 

"Cf.ffijEWH^S^'Si, by M& : Ch ' ao ch ' in 9 ch ' ao W ^i 

86 Cf. H^^flRfi^JliWJfiff , by Hu Shih, op, ciU 95-102. 



Diagram IV ] 









Kao tsu 

Tseng tsu 

Tseng po 

Tseng shu 


Mourning Grades: 
1 = ch'i, one year 
% = Ta kung, nine months 

3 = Hsiao kung, six months 

4 = Ssu ma, three months 

I, II, III, IV indicate generations removed from ego 


Shu tsu 

Tang po 

Tang shu 





Tang po 
Tang shu 

Tsai ts'ung 


Tsai ts'ung 




Tang hsiung 
Tang ti 

Tsai ts'ung 


Tsai ts'ung 


Tsu hsiung 


4 ^=i 







Tang chih 

Tsai ts'ung 

Tseng sun 

Chih sun 

Tang chih 

Tseng chih 

Hsiian sun 

1 Since the author is merely outlining the basic mourning system, only 
the male relatives are given. A complete specification would require 
eight to twelve diagrams. 


ceremonial and social life. Under their influence the Chinese kin- 
ship system, through the increasing emphasis laid upon collateral 
differentiation and generation stratification, was transformed from 
a classificatory system based upon an exogamous sib organization 
into one of the descriptive type. 37 

Cross-cousin Marriages 

Cross-cousin marriage is permitted, but not encouraged, in 
modern China. Generally, it is discountenanced not on the ground 
that the blood relationship is too close but on the ground that 
older relatives might be estranged as a result of difficulties which 
might arise between the young married couple, or vice versa. On 
the other hand, however, it is desirable, because it increases the 
number of relationships and knits the bond more closely. 38 
Theoretically and ritually, it has been disapproved since the be- 
ginning of the first century A. D. 39 Legal prohibition, however, 
came rather late, the first definite clause being found in the Ming 
Code. 40 Since the enforcement of this law proved rather difficult, 

87 Kingsley Davis and W. Lloyd Warner have made some very pertinent remarks 
concerning the use of " classificatory " and " descriptive H in connection with kinship 
analysis (Structural Analysis of Kinship, American Anthropologist, 39: 2 [1937], 291- 
315) , They have also formulated a new set of categories for the structural analysis 
of kinship, with which, I think, few students will agree. The present writer disagrees 
with them in many points on their interpretation of the Chinese system. Since, how- 
ever, this MS is going to press as their article appears, it is not possible to elaborate 
this remark. 

88 As the popular saying goes, il_hJJDiS. 

**Pai hu t'ung, 10.16a: ^M4^JlB ± , #^f# 3&fc . ^1*1$0, %$ 
#!£-&. Jt^j of the Chin dynasty, says, [Tung Tien, 60. 16b-17a, ft^gvpRf j$ 

nm h:^aw*«», mm b: +*;£«, a#Hfo», wmsLm 

II^RTi§» itb^P^nl^^-ui, These statements probably read too much into the 
old ritual works. 

"Ming lu chi chieh, 6.17a: ig^fi;£M*M»J*#, &A+0MJI. 
The clause in the Tang Code [Tang lu shu 1.14. 2b], ^3t#]£^M^^fr#|c 

• • • Sfe^^^^W • • • is sometimes expanded to include prohibition against 
cross-cousin marriage. Cf. Jung chai sui pi (hsil pi) , 8. 12. The clause seems to indicate 
only parents' cross-cousins, not one's own cross-cousins; if so, the interdiction is against 

inter-generation marriage rather than cross-cousin marriage. But the T'ung tien (95. 11a) 
seems to show that during the T'ang period marriage with cross-cousins and mother's 
sister's daughter was actually prohibited. 


in the Ch'ing Code this interdiction was invalidated by another 
clause, immediately following it, which allowed such marriages. 41 
It must be noted that in modern China the kinship system is not 
the primary regulator of marriage; the important factor is sib 
exogamy supplemented by the generation principle. Thus, not 
only marriage with cross-cousin, but also with parallel cousin by 
mother's sister is allowed. 42 No statistical data for cross-cousin 
marriage are available at present, but my general impression is 
that the percentage is very small. In any event, cross-cousin 
marriage is in no way reflected in the modern kinship system. 43 
The ancient system as recorded in the Erh Ya and the / Li 

41 G. Jamieson: Translations from the General Code of laws of the Chinese Empire, 
Chapter 18: "A man cannot marry the children of his aunt on the father's side, or 
of his uncle or aunt on the mother's side, because though of the same generation they 
are within the fifth degree of mourning." But a little later in the Li, it reads: "... 
In the interest of the people it is permitted to marry with the children of a paternal 
aunt or of a maternal uncle or aunt." China Review 10 (1881-82), 83. Cf. also Sir 
G. T. Staunton, Ta Tsing Leu Lee (1810), 115. 

42 The practice of marriage with the mother's sister's daughter began at least as 
early as the third and fourth centuries A. D. E. g., Tung Tien 95. 9a cites the dis- 
cussion of ^^ of the Chin dynasty concerning the mourning obligations of dual 
relationship for mother's sister's daughter who married one's own older brother. A 
great many useless discussions have been lavished on this subject. 

43 T. S. Chen and J. K. Shrtock in their " Chinese Relationship Terms," American 
Anthropologist 34: 4, 623-669, interpret in terms of cross-cousin marriage the fact that the 
father's sister's children and the mother's brother's children are designated by the same 
terms (see Chen-Shryock Table I, terms 85-92, and Table IV, terms 17-24, and note 33) . 
But I do not see where the marriage element enters. The terms merely indicate cross- 
cousinship and nothing more. In order for these terms to be interpretable in terms of 
cross-cousin marriage, either the mother's brother or father's sister's husband must be 
addressed by the same term as that used for wife's father, or sister s husband or wife's 
brother addressed in cross-cousin terminology — in fact, any usage that will bring in the 
marriage element. Unfortunately, no terminology of this sort exists in the modern sys- 
tem. Hence, the authors' interpretations involving cross-cousin marriage in notes 33, 34, 
39,42,61,64,65 and 67, are untenable. Furthermore, these interpretations are based on 
incomplete and faulty data. E. g., the important modifier piao is omitted from the terms 
of mother's sister's children, thus making mother's sister's daughters merge with wife's 
sisters. Mother's sister's children are piao, just as mother's brother's children and 
father's sister's children are piao. Not only cross-cousins are designated by piao, but 
also parallel cousins through mother's sister. This consideration completely invalidates 
the cross-cousin interpretation. Actually, since, as the authors have ably shown, the 
abandonment of the cross-cousin marriage custom was responsible for the development 
of the modern system, how can the modern system still be interpreted as indicative 
of that usage? 


reflects a preferential type of cross-cousin marriage in certain 
kinship equations. 44 The terms concerned are the following: 45 
Chiu Jt : a. mother's brother, b. husband's father, 46 c. wife's father, 
as wai chiu* 7 Ku J£: a. father's sister, b. husband's mother, c. 
wife's mother, as wai ku. Sheng i%: a. father's sister's sons, b. 
mother's brother's sons, c. wife's brother, 48 d. sister's husband 
(man speaking) , 49 These terms indubitably manifest cross-cousin 
marriage of the bilateral type, coupled with sister exchange. The 
latter practice is shown especially in the term sheng, which means 
wife's brother and sister's husband. 

Indirect evidence can be obtained from the arrangement of 
the Erh Ya. Here the terms on kinship are arranged into four 
groups: i. Tsung tsu mM, Relatives through father, ii. Mu tang 

14 The first to interpret the Erh Ya system in terms of cross-cousin marriage was 
M. Granet, La civilization chinoise, 187, The thesis was further developed by Chen 
and Shryock, op. ciL 629-630. 

" In the following notes some of the old Chinese interpretations of the extensions of 
these terms are given. They are not necessarily correct, but they do serve to show 
the traditional Chinese conceptions. 

46 The old interpretation of the extension of chiu and ku to include husband's father 
and mother is as follows: The one who is as venerable as father, but who is not the 
father, is mother's brother, chiu. The one to whom one is as attached as much as to 
mother, but who is not the mother, is father's sister, ku. Husband's parents are of 
similar relationship, hence we call them chiu and ku. Cf. Pai hu t'ung, 8 20b- ff|-^c 

^^#fiMMH? #*n£ilff#£#,JUlL. *ito#ro*«*,«r4lL. tit 

47 Wife's parents are called wai chiu (outside chiu) and wai ku (outside ku) , that 
is to say, the wife is an outsider who comes to one's own family and makes it her 
own family too. She calls the husband's parents chiu and ku. The husband, in 
reciprocating, calls her parents wai chiu and ftoai ku, as a sign of equality for both 
parties. Shih Ming: *2«BftM. #B ft**, £¥*£&#, ffl 3! BSU& 

48 Wife's brothers are called wai sheng (outside sheng) because their sister marries 
ego and becomes ego's wife, hence her male siblings are the sheng of an outside sib. 
Shih Ming: #£ JljfcHftiJ , ^m^^Ml , £jff SftJSS, &$:%%& 

49 Kuo P'o's commentary on the term sheng, in the Erh Ya, says that these four 
individuals are of equal status, hence they reciprocally call one another sheng. HA 
fifllSt* SfcBl+BJSSJ. This conception of f§5tff|, equal status, between the relation- 
ships of sheng, has been the basis of most later interpretations; e. g., ^fi^-f^SS? 5 


#l£, Relatives through mother, iii. Ch'i tang 3Hf£, Relatives 
through wife. iv. Hun yin MM, Relatives through husband. 
It is interesting to note that father's sister's sons, mother's brother's 
sons, sister's husband (M. S.) , and sister's sons are all listed under 
Group iii, ch'i tang. The grouping of these relatives of quite 
distinct affiliations under Relatives through wife clearly demon- 
strates that the Erh Ya system is built upon the practice of cross- 
cousin marriage. 


The sororate was operative during the feudal period, at least 
among the feudal lords. 50 The Erh Ya gives the term for sister's 
husband [woman speaking] as ssu, literally " private." This is 
sometimes interpreted as evidence of the sororate, since a woman 
considers her sister's husband her " private." 51 The validity of 
this reasoning is rather dubious, without other terminological 
corroboration. Ts'ung mu, the term in the Erh Ya and the I Li 
for mother's sister, has also been interpreted as a reflection of 
this usage. Ts'ung mu literally means " following mother." 52 
But the term is best interpreted as a counterpart of ts'ung fu, a 
term for father's brother; hence ts'ung indicates the collateral 
line rather than potential motherhood. 

Throughout the historical period, including modern China, the 
sororate has been practiced, but, since the possibilities are reduced 
by infant betrothal, its occurrence has been sporadic only. The 

60 Cf. M. Granet, La polygynie sororale et le sororat dans la Chine jeodale, 1920. 
The author overworks his material to arrive at forced conclusions, but most of the 
relevant data are collected in this little work. The thesis is also incorporated in his 
later work, La civilization chinoise, in which he has utilized the antiquated anthropo- 
logical theory that the sororate and levirate represent survivals of an early fraternal 
group marriage. 

51 Cf. Chen and Shryock, op. eit., 628. The Shih Ming interprets ssu in quite a 
different way. Female ego's sister's husband is called ssu, private, because this man 
has private relations with female ego's sister: ^ifl^S^Bpi^H^t * HfJfcfeWi 
Wt'iLty ? lHlAJ£| ^jffoffcf] J&MtJftL • On the basis of this old interpretation, the 
ssu does not concern female ego at all, and so is scarcely evidence for the sororate. 

B2 Liu Hsi (ca. 200 A. D.) , in his Shih Ming, interprets tsung mu in the following 
manner: mother's sisters come to marry the father as ti $[?, hence they are of the 
status of tsung mu. Even if they do not marry the father, the term is still applied 
to them. This, I believe, is the earliest known sociological interpretation of kinship 
terminology. Whether or not it is correct is quite another matter. 


single reflection in the modern system comes in the term i, which 
means mother's sister, father's concubine, wife's sister, and con- 
cubine. This identification seems irrefutable, yet another ex- 
planation is possible. A man calls his friends hsiung, older brother, 
and ti, younger brother, as a sign of courtesy and intimacy. A 
woman also calls her female friends tzu, older sister, and met, 
younger sister, for the same reason. It is perfectly natural for 
the wife to call and consider her husband's concubines mei 9 and 
actually she does. The term i may thus be extended without 
recourse to the actual sororate at all. Similarly, i is usually used 
by children for the father's concubines, 53 and by servants for the 
master's concubines, in both instances as a complimentary term. 

Where the sororate is practiced extensively, it may be accom- 
panied by marriage with the wife's brother's daughter, because 
if the wife has no marriageable sister her brother's daughter is 
a good substitute. There are indications of such a practice in 
feudal China among the nobility. 

When a feudal lord married, his bride was accompanied by 
eight bridesmaids called yin, who were his future concubines. 54 
The yin were recruited in the following manner. The bride and 
the eight yin were divided into three groups, with three women 
in each group. The first group consisted of the bride, one of her 
younger sisters or younger half-sisters, ti, 55 and one of her older 

53 The Tung su p'ien [18. 17a-17b] states that the father's concubines are called t, 
because of the old yin marriage custom. / was originally a term for several sisters 
who married the same husband. In later times, the yin custom was discontinued but 
concubines are actually equivalent to the yin. Therefore, although father's concubines 
are not mother's sisters, i can still be applied to them. This sociological interpretation 
is rather erroneous. The yin marriage custom had nothing to do with the term i, and 
the yin's were never called i, but always chih ^ and ti $jjp . 

"Kung-yang chuan ^^+A^. 8. lb-2a : J$:#fPj? $&?£— ®,n®ft 

65 Whether ti $$ meant the bride's actual younger sisters, or her younger half- 
sisters, i.e. her father's yin's daughters, is a matter of speculation. Probably ti meant 
only her younger half-sisters, since there is abundant evidence to show that the 
daughters of the principal wife, fu jen t were always married out as ju jen, and yin's 
daughters always married out as yin. In this way the noble-born and low-born were 
always kept constant jl| j§l^=f 'ft*' . This interpretation tallies with the uses of i and 
ti in the Erh Fa. Cf. ^[^#: Shuo wen chieh tzu shu i Ift^CM^^St 9 (M^ 
$?!#*), 2.44a. 


brother's daughters, chih. These three women constituted the 
principal group. Two other feudal states of the same sibname as 
the bride each supplied a principal yin, a ti, and a chih™ Thus 
there were three groups and nine women in all. The contribution 
to the yin by other states had to be entirely voluntary, and could 
not be solicited, 57 for it was not proper to ask children of others 
to become the dishonorable yin. 58 

This elaborate system for the selection of the yin served to 
insure a large number of descendants for succession in the feudal 
lord's office. 59 A niece, rather than a second younger sister was 
included in the yin in order to create a difference in the blood, 
so that if the two sisters failed to bear issue, a niece of different 
blood might bear a son. Similarly, the two other groups of yin 
were selected from two different states, in order that their blood 
would be still more dissimilar and the chances of having an heir 
would thus be tripled. 60 Sib relatives were selected for the yin 
with a view to preventing jealousy and intrigue within the harem. 61 

The yin custom was not strictly what is usually termed 
"secondary marriages." AH the women were married at once, 
though if a yin were too young she could " wait her years in her 
parental state " until grown up, and was then sent to the bride- 
groom; 62 but this was rather unusual. In theory, a feudal lord 

■T«> Chuan Jjft^-hA^, 26.23a: *A#«*tE, H-tii . /Lf*&#:fc, 

IPIttJfifcfc, MfflEfDi^* This may have been the general rule, but there were 

"Kung-yang chuan &&+.*,$£, 8.2a: |SJflC&: WttJEfct^t, H. ^3& 

" Pai hu t'ung, 10. 12a: 0f J&*3»*#, fSJ? AW« , «fc#^ , fc^ 

#A®&& • ■ . MJt^ftgtf. 

"Ibid., 10.11a: 3^f$«— 3yL£*ffl? mmWMM. Ibid., 13b, * 

«fi££f, #fllA£#, fiHJKiML. 

" Pai hut'ung, 10.11b: ^M^ffl? tfH^Ut. SH^cM? MMM 

"ibid., if &$#£#, «#^^+B**54fc. -AW, HA&£, ^a 

&£&. Kung-yang chuan, ffifttA*. 8.2a: fpjflc&: &#&«&&;£#, 

«ffi-Aw,-A#aL. mmmm, ^mmm^i. mmttm 

"Kung-yang chuan %$&*. 3.8a: &%££)&£. ft: &&&* fG$E£J& 


married only once in his life ; 63 if the principal wife, fu jen, died 
a yin might act for the fu jen in her ceremonial and social capaci- 
ties but could not assume the title of fu jen. 6 * Actually, the yin 
had very little legal status, and were only " legalized mistresses M 
of the feudal lord. 

The ministers, ch'ing ta fu, of the feudal lords could have one 
wife and two concubines, but could not take the wife's sisters or 
nieces as yin 65 There are discrepancies in the interpretation of 
this rule in the classical works, and in a few instances the ministers 
followed the feudal lords' example and took the wife's younger 
sister and niece as concubines. 60 The scholar class, shih, i. e., the 
lower ruling class, were allowed one wife and one concubine. 67 
Opinions differ as to whether or not this one concubine could be 
the wife's younger sister or niece. According to Ku Yen-wu 
(1613-1682) , a scholar could not take his wife's younger sister or 
niece as concubine. 68 All commoners were allowed only one woman, 
at least in theory. 69 

To judge from the evidence, the yin custom may have been 
only a kind of " legalized incest," whereby the emperor and feudal 
lords might assure themselves of an heir. Some of the lower 
nobility, however, might have followed suit later. It is not difficult 
to see why such a highly arbitrary custom could not have be- 
come very prevalent even among the nobility; not only was the 
supply of women limited, but the practice actually ran counter 
to the generation-principle ideology of this period. For example, 

88 Ibid., ££H^l^> 8.2a: f&£— US*.*, f&fi^M. 

" Tso Chuan gf^Tt^, 2. 2b-3a: SiP* , fltSJ^iP. tfcft: f&fl|*&3fc, 

5jf This rule was not absolute, cf. Pat hu t'ung, 10. 18-19. 
"Paihut'ung, 10. 17b: »■*:=£--#- *ft? MStlftiMiL. ^fllStt 

" E. g., Tso Chuan, 85. 18.: fcWKgfc^flHBr , ^MR^M^ , H^ ##«. 
"Pai hu fung, 10.18a: ±— #— ^M? T^PA^lM. 36J&/M&S •' ± 

"Jih chih lu, 5. 84b: -f|;g> H^cf&. 

"This is what was called [S^|ZE$S- In fact, under the feudal system, the title 
to all land was held by the feudal lord, and the common people worked under a 
" serf " system; no one could afford two women unless he belonged to the ruling class. 


a feudal lord was not allowed to marry the noble women of his 
own estate, because, theoretically, everyone within his feudal state 
was his subject, and if the lord were married to any woman in 
his own state her parents would automatically be a generation 
higher than he and thus could no longer be his subjects. To avoid 
this contradiction, a feudal lord was required to marry outside 
his own state. 70 Yin marriage was swept away with the feudal 
system during the third century B. C. Since the beginning of 
the Western Han period [B. C. 206-A. D. 8], the practice has 
never again been recorded, either among the royalty or the 
nobility. 71 

When we turn to the ancient kinship system, we find there a 
peculiarity which seems to reflect the practice of marriage with 
the wife's brother's daughter. In the Erh Ya sister's sons are 
called chu (M. S.) . In a later passage, the same relative is called 
sheng. It has been remarked above, in connection with cross- 
cousin marriage, that in the Erh Ya, sheng is principally used for 
father's sister's sons, mother's brother's sons, wife's brothers, and 
sister's husband (M. S.) . This use of sheng to mean sister's sons 
does not comply very well with the generation principle stressed 
in the Erh Ya. 12 

There is also a peculiar usage of sheng in the works of Mencius 
[B. C. 373-289]. Mencius used sheng to mean daughter's hus- 
band. 73 It seems that the overriding of the generation principle 
in the use of sheng, was a phenomenon that appeared rather late 
in the feudal period. Sheng was applied, during the feudal period, 
to 1. father's sister's sons; 2. mother's brother's sons; 3. wife's 
brothers; 4. sister's husband (M.S.) ; 5. Sister's son (M.S.) ; 6. 
daughter's husband. 

The first four connotations can be interpreted in terms of cross- 
cousin marriage of the bilateral type, together with sister ex- 

"Paihufung, 10.15a: f^^J^^g ^M +M ? ftf£Wf##, H 

71 Some students have even suspected that the yin marriage custom was a mere 
invention of the Han scholars. 

72 The only other term is the Erh Ya which overrides generations, is shujffli, for 
father's younger brother and husband's younger brother. 

"Ming tzu, 10A. 10b: ^ffijllfr, ^ft^J^ .... 


change. The last two meanings seem to demand a cross-cousin 
marriage of the above type, together with a marriage with the 
wife's brother's daughter. In a case of this sort, both sister's 
husband and sister's son can marry ego's daughter, and ego's 
daughter's husband will be identified with both sister's husband 
and sister's son. However, in view of the fact that marriage with 
the wife's brother's daughter was only a " legalized incest " among 
the nobility and never a prevalent practice, teknonymy is a more 
plausible, and a simpler, explanation. Both sister's son and 
daughter's husband will be sheng to ego's own son, if cross-cousin 
marriage is assumed ; ego simply adopts the son's term in address- 
ing them. 

On the whole, the influence of the sororate, both on the ancient 
and modern kinship systems, has been rather negligible. Insofar 
as the evidence goes, the sororate, both in ancient and modern 
China, is only a permissive type of marriage, that is to say, ego's 
marriage with one woman does not affect the status of marriage 
of her sisters, nor does it affect ego's own marriage status. 

Levirate 74 

The junior levirate certainly exists in a few parts of modern 
China, at least among the poorer classes/ 5 but, even in the few 
places where it is practiced, it is not considered respectable. A 
man adopts this only as a last resort in getting a wife. If necessary, 
he can sell his brother's widow and use the " bride-price " to 

74 Sir James Frazer has insisted upon the intimate co-existence of the sororate and 
the levirate (Totemism and Exogamy, 1910, 4.139-150). R. H. Lowie also says: 
" The connection would undoubtedly appear to be even closer were not much of 
our information on marriage rules of primitive tribes of rather haphazard character. 
That is, it may safely be assumed that in not a few instances it is sheer negligence or 
defective observation that has made writers report one of the two customs without 
the other." (Primitive Society, 1920, p. 36.) If this correlation is valid, we should find 
the levirate in China as a correlative institution. 

7B Cf. China Review, 10 [1881-2], 71, The levirate in China. Also, Huang Hua-chieh 
%^W> Shu chieh sao $L$£%1, Eastern Miscellany, 31: 7 [1934], #§ 20-21. P. G. 
von Mollendorff once remarked, " I have not been able to find the slightest trace 
of it (levirate) , and it can never be of the same importance with the Chinese as with 
other people (e. g. to keep the family property) , as posthumous adoption, the Chinese 
substitute for it, fully meets the object. The Family Law of the Chinese [1896], 17. 


marry another woman. Legally, marriage with the older brother's 
or younger brother's widow is stringently prohibited ; 7e the pun- 
ishment is strangulation for both parties. G. Jamieson has doubted 
its existence in China at all under such heavy penalties. 77 

Whether or not the junior levirate existed in ancient China 
is quite problematical. Granet cites two cases from the Tso 
Chuan, but these can hardly be interpreted as evidence for the 
levirate. 78 Chen and Shryock say that " the relationship terms 
indicates only the junior levirate, in which an older brother mar- 
ries his deceased younger brother's wife. A wife calls her husband 

76 This law was first explicitly stated in the Ming lii chi chiek 6. 20, promulgated 
during the period Hung-wu, 1368-1398 A. D. (latest revision, 1610 A. D.) . In all 
earlier codes the levirate was prohibited under a more general clause, e g in the 
Tang lii shu i,li.S>: H*#4flft*£#]ffi#«*, ##— W. «jftftJ! 

»#.#— ¥. A<$lB±.M&lk. %&m~m, E*l£. : Jnder this clause, 
the levirate is out of the question. 

On the other hand, the explicit clause in the Ming Code against the levirate may 
be a reaction to Hs introduction into China through the Mongols of the Yuan dynasty. 

No. 8, April 10, 1936, (Sheet 11), p. 3. 

77 China Review 10. 83 says: " In view of the severe penalty for it, it is scarcely 
possible that the levirate can be practiced in any part of China." 

78 La civilisation chinoise, 424-5. The two cases cited are: Pi Wu-ts'un [Legge, 
Chinese Classics, V: ii, 773]; Tzu Yuan and Hsi Kuei [Ibid., V: i, 115]. In the first 
case, Pi Wu-ts'un, who was going to war and intended to marry a woman of better 
status, refused his father's proposal on the pretext that his father would be able to 
marry his younger brother to the woman in question. The woman proposed by his 
father not only cannot be regarded as his wife, but not even as his fiancee. In the 
second case, Tzu Yuan already had the full authority of the state of Ch'u. and did 
not need to marry his older brother's widow in order to acquire his brother's authority, 
as Granet's theory demands. On the other hand, Hsi Kuei was a noted beauty and 
Tzu Ytian's attempted seduction was motivated by lust. Unfortunately for Granet's 
thesis, Tzu Yuan did not succeed in seducing her, and was soon killed. 

According to the ancient mourning specifications, sister-in-law and brother-in-law 
are not subject to mourning obligations, hence Granet considers his hypothesis con- 
firmed inasmuch as relatives without mourning obligations may marry. But mere 
mourning obligations do not prevent marriage, e. g., cross-cousins have mourning 
obligations to each other but, according to Granet's theory, they are prescribed to 
marry ! 

What we need are actual instances of the levirate — not this ambiguous and anomalous 
kind of material which may be interpreted to support any kind of hypothesis. Granet 
is laboring to use the sororate and levirate as proof of an earlier fraternal group- 
marriage which is in itself a hopeless hypothesis. 


and her husband's older brother (possible husbands) by the same 
term, po, but uses a different term for his younger brother." 78 
The authors are somewhat confused in this. Where only the older 
brother can marry the younger brother's widow, the practice is 
termed the " senior " levirate, not the " junior " levirate. The 
" senior " levirate alone is not found in Asia. Either both forms 
of the levirate are practiced by the same people, or the junior 
levirate alone is practiced, as in India, southeastern Asia, and 
northeastern Asia. There is also an anachronism in the connota- 
tions of po cited by the authors. In a few places in the " Book of 
Odes," po is interpreted as meaning husband, 80 but this usage is 
not found in literature after 500 B. C. On the other hand, the 
use of po to mean husband's older brother did not begin until 
the tenth century A. D. 81 Thus, po meaning husband and po 
meaning husband's older brother not only are not contemporane- 
ous, but are separated by a hiatus of fifteen centuries! Historically 
minded though the Chinese may be, I do not see how one can 
interpret this terminology in terms of the levirate. 

Chattopadhyay has interpreted the differentiation of older and 
younger brothers in India in terms of the junior levirate. 82 Such 
an interpretation is extremely weak, unless supported by other 
terminological corroboration. If the junior levirate explains the 
differentiation of older and younger brothers, it certainly does not 
explain the differentiation of older and younger sisters, which is 

79 Chen and Shrtock, op. cit t 628-29. 

*°Shih Ching, 3C. 7a-8b: fg^ig^, ^±^^1 f^ffe^ ,ft3Eff!B. g 

fcjft, "MrAtfRah gftHItt, t£S«£? • ■ ■ ; The po used here k some- 
times interpreted as meaning husband, but it is uncertain whether po is a relationship 
term for husband, or a reference to the official title the husband holds, or simply a 
word meaning " the brave and handsome one." To judge from the context, the last 
is the preferable interpretation. 

"T'ao Yo Kf-ft, Wu tat shih pu 5ftM (!Mt$#*)» « ■ 8a: BO * 

82 Chattopadhyay, Levirate and Kinship in India, Man, 22 (1922), 25. W. Lloyd 
Warner, Kinship Morphology of Forty-one North Australian Tribes, American Anthro- 
pologist 35. 66 makes a similar interpretation. 


also characteristic of most Indian systems. Whether or not the 
tsung fa adequately explains the expression of the category of age 
in generation in the Chinese system, as advocated by the old 
Chinese authors, we do not know; certainly this expression can- 
not be explained by the junior levirate, which is of so sporadic 
occurrence in China. 


We have already discussed several types of marriages that are 
relevant to the determination of minor kinship peculiarities. With 
the exception of cross-cousin marriage, the influence of the others, 
both on the ancient and modern systems, has been rather negligible. 
These problems will be further discussed here in connection with 
teknonymy and other peculiarities in the system, in order to 
ascertain the actual determining factor or factors. 

As has already been shown, generation is an important struc- 
tural principle in the Chinese system. It also regulates marriage 
among relatives, and plays an important role in the functioning of 
Chinese social and ceremonial life as a whole, since the dealings be- 
tween relatives are in many respects based upon generation differ- 
ences, so also in the assignment of mourning grades, etc. Since 
generation is such an important factor in the system, we should 
expect it to be consistently expressed in terminology. Yet there are 
some notable exceptions. Mother's brother and wife's brother are 
designated by the same term, chiu\ mother's sister and wife's 
sister by the same term, i; father's older brother and husband's 
older brother by the same term, po ; father's younger brother and 
husband's younger brother by the same term, shu ; father's sister 
and husband's sister by the same term, lew, etc. These peculiari- 
ties are of significance, because originally the generations of these 
relatives were clearly differentiated by distinct terms, and only in 
the course of time were they gradually merged into each other. 

First, let us take the connotations of the term chiu, and the 
terms for the wife's brother, during the various periods, and 
arrange them in a single table, as follows: 


Period Connotations of chiu Terms for wife's brother 

I mother's brother 

1st Millennium B. C. husband's father sheng 

wife's father 

1st Millennium A. D. mother's brother fu hsiung ti 

III mother's brother 

2nd Millennium A. D. wife's brother chiu 

The various connotations of the term chiu in Period I are perfectly 
intelligible from the point of view of cross-cousin marriage, as 
discussed above. In such a marriage, the mother's brother and 
husband's father is the same person, so also the mother's brother 
and wife's father. In Period II the cross-cousin marriage was 
dropped, and consequently the meaning of chiu became confined 
to mother's brother. 

The terms for the wife's brother are different for each of the 
three periods. In Period I wife's brother was called sheng® . Sheng 
also meant, in this period, father's sister's son, mother's brother's 
son and sister's husband (man speaking) . 83 This is also explicable 
in terms of cross-cousin marriage of the bilateral type, coupled 
with sister exchange. In Period II, because of the disappearance 
of this type of marriage, sheng was no longer applicable to any 
of these relatives and new terms were introduced to take its places. 
Fu hsiung ti is the term used for wife's brother. 84 

In Period III the term chiu (mother's brother) was extended 
to include wife's brother. The first use of chiu in this new mean- 
ing is to be found in the Hsin Tang Shu, In the biography of 
Chu Yen-shou, we read: " Yang Hsing-mi's wife is the older sister 
of Chu Yen-shou. . . . Hsing-mi [luring Chu Yen-shou into a 

88 Cf . Erh Ya. 

84 The term is purely descriptive. Fu j$ means wife, hsiung ti means brother (older 

and younger) . Pei Ch'i Shu <S^{$, SO. 9b-10a: H^ j£ g , f%^^±> #§5t O 
^C» MPiS- ibid.. ggTClItt, 29.6b: ffi^jO^c, j|***l %. CKihsiung 
ti ^JL^ti and net hsiung ti P}5t^6 were also permissible at this time. Cf. Table III. 
[From this point the editors have been forced to substitute the symbol O for some 
frequently recurring characters. In this instance the missing characters are No. 11 
on p. 149.] 


trap] says, * I have lost my eyesight and my sons are too young. 
Having chiu [meaning Chu Yen-shou] take my place, I shall have 
no worry/ " 85 This is certainly a curious extension of the use 
of chiu, for, through all the previous vicissitudes of the term, the 
generation element had always been preserved. This overriding 
of the generation principle certainly warrants an explanation. 

A strictly sociological interpretation would point to a marriage 
with the wife's brother's daughter. In Period I chiu also meant 
father-in-law; since, in this interpretation, the wife's brother is 
a potential father-in-law, the extension of the term chiu to in- 
clude him would be perfectly logical. However, there are several 
difficulties in such an interpretation. In the first place, historical 
evidence does not seem to support this hypothesis. Wife's brother's 
daughter marriage in connection with the yin custom was never 
a preferred form, nor, as stated above, was it common even among 
the feudal nobility. Moreover, it disappeared together with the 
feudal system during the third century B. C, and has never been 
practiced since. Secondly, chiu had ceased to mean wife's father 
at least a thousand years before it was extended to mean wife's 
brother. These two temporal considerations, involving a hiatus 
of more than a millennium, are irreconcilable with such an inter- 
pretation. Thirdly, such a marriage contradicts the generation 
principle ; wife's brother's daughter is one generation lower than 
ego ; and thus, in the Chinese system, is within the incest group. 
Legally, inter-generation marriage between all relatives became 
definitely prohibited at least half a millennium before chiu was 
extended to mean wife's brother. 86 In the face of these objections, 
the above interpretation is untenable. 

It is significant that Chinese scholars had been employing 
teknonymy to explain this terminological anomaly long before 
the introduction of the term into anthropological discussion by 

**Hsin Tang Shu £jf«ffi. 189. 10a: fflSKlBHkJfe, M&lftB : &^ 

mn^mnm. mm. -Ammfft. *?#*£, *&^ a, mw*. 
ii. $, &s*sMb, #2. ft%m&: ©«bj!, ttiP#r, #0fe, imm 

■^. [The missing character is No. 16 on p. 149.] 

"The T'ang Code, compiled and promulgated during the period A. D. 627-683, 
stringently prohibits inter-generation marriage: Tang lii shu i 13. 2. 


E. B. Tylor. 87 Ch'ien Ta-hsin [1727-1804], one of the most pene- 
trating classical scholars of his time, attributed this extension of 
the meaning of chiu to the gradual, imperceptible effect of the 
practice of teknonymy. 88 Wife's brothers are chiu to one's own 
children. The father, adopting the language of his children, also 
calls his wife's brothers chiu. This process can clearly be seen 
in the above-mentioned instance of Chu Yen-shou. Yang Hsing- 
mi called Chu Yen-shou chiu, at the same time mentioning his 
own sons. One can infer that, after long teknonymous usage, the 
term chiu established itself and finally displaced the older term. 

Whether or not this hypothesis can be sustained depends upon 
the additional evidence we can adduce for its support, or, in other 
words, upon whether or not it can explain all the peculiarities of 
the same nature in the system. Let us now turn to the examina- 
tion of those terms by which the wife addresses her husband's 
brothers: po for the husband's older brother, and shu for the 
husband's younger brother. But po was originally a term for 
father's older brother, and shu for father's younger brother. This 
overriding of generations is quite aberrant, from any point of 
view. Insofar as I am aware, there is no social or marital usage 
in China, nor is there any comparable usage that ethnographic 
data suggest, which could give rise to such a terminology. 

From the historical point of view, the terms for these relatives 
were different at different periods. In the Erh Ya the father's 
older brother is called shih fu. From the second century B. C. 
down to the present po fu has been the standard term, but from 
the fourth century A. D. on, po alone also has been in use. 89 

In the Erh Ya husband's older brother is called hsiung kung. 90 
During the succeeding centuries hsiung chang was commonly em- 

87 On a Method of Investigating the Development of Institutions .... Journal, 
Anthropological Institute, 18 (1889), 245-69. 
"Heng yen In, 3. 18b: ^Hjf • ■ ■ ««*£»»»»*« , «*?££ 

"Li Chi H-ipB, 18. 10a: g#ffif.m&3&&3£- Y <™ 'hih chia hsiin 
i. 6b: ^A%VM&'£%.'£ , mWk&WWbU . Cf. Heng yen lu, 3. 6-7. 

State University of New Yorh 

Stony Brook 

I IR0A.B!F<; 


ployed. 91 Ca. the tenth century A. D. po was extended to include 
husband's older brother. 92 

In the Erh Ya the father's younger brother is called shu fu. 
This term has continued in use without any radical change down 
to the present ; as in the case of po, above, from the fourth century 
A. D. on shu alone has also been in use. Shu is also used in the 
Erh Ya for husband's younger brother. This usage is rather un- 
usual, inasmuch as it overrides generations and thus contradicts 
its own statement of principle, that " husband's siblings are affinal 
siblings." 93 

As has been stated above, no possible explanation for this 
blending of generations can be found in marriage forms ; the only 
possible alternative is teknonymy. Husband's brother's are po 
and shu to female ego's own children. The mother, adopting the 
terminology of her children, also calls them po and shu. This case 
tremendously strengthens our hypothesis, since no other known 
social factor or form of marital relationship can adequately explain 
these usages. 

A similar situation exists in the terminology for father's sister 
and husband's sister, both called ku, and mother's sister and 
wife's sister, called i. 

As has already been pointed out, ku is used in the Elrh Ya for 
father's sister, husband's mother, and wife's mother [e. g., wax ku], 
as a result of cross-cousin marriage. When cross-cousin marriage 
declined, ku was usually employed for father's sister alone. In 
the Erh Ya, husband's older sister is called nu kung, and younger 
sister nu mei. 94. Somewhat later shu met was used for the hus- 
band's younger sister. 95 In the fourth century A. D. the term ku 
began to be extended to include husband's sisters. 96 The factors 

91 Shih Ming: JcZR . . . f$fJ8Bj£$. Cf. Table IV. term 5. 
"See p. 193, note 81. 

"Erh Ya: WZ&®]l$R% , 0£fltlUfl Jt*. [The missing character is 
No. 18 on p. 150.] 

94 ^-ZMiMiC^y^-ZiK^^-kO . [No. ll on p. 149.] 

"Hon Han Shu WW# DM 4fc 1* 8b: ^A£#«*±, * O 

*£;£$f £-& ! O #f;£ J& B , &M ;£# B& . [First and second are No. 16; 

last, No. 11 on p. 149.] 

Be The earliest occurrence of the term hsiao hi for husband's sister is in the famous 


behind this extension cannot be exactly ascertained, although the 
marriage rules and social customs of the period concerned are 
fairly well known. The extension could not have been due to mar- 
riage with the wife's brother's daughter, in which case the hus- 
band's sister would be elevated to the position of the husband's 
father's sister ; the above-cited objections to a similar interpreta- 
tion of chiu also apply here. Furthermore, other features do not 
follow either terminologically 97 or conceptually. 98 Teknonymy 
remains the best explanation since husband's sisters are ku to 
female ego's own children. 

Originally i was used, as in the Erh Ya, for wife's sisters. In the 
Erh Ya mother's sisters are called ts'ung mu. The earliest use of i 
to mean mother's sister is found in the Tso Chuan. In the twenty- 
third year [B.C. 550] of Duke Hsiang, a passage reads, " Z's 
daughter of Mu Chiang." 99 By checking the relatives connected 
with Mu Chiang, one finds that the term, i, here does not mean 
wife's sister [or married sister, woman speaking] as it should, but 
mother's sister. As a matter of fact, the passage should read, 
" ts'ung mw's daughter of Mu Chiang," not, " i's daughter." 

Theoretically, the sororate, together with a marriage with the 
father's widows, would adequately explain the usage. In this com- 
bined type of marriage, the father marries mother's sisters and 

poem ■£SI$Siift'Jtt#ft £ ( F * '*» *«*» V un 9 chi, 1- 17b): *PJ|/h##], KS 

SfcftiP. *f#&J#P¥, /M**6ife#, -*HiW«, M$tom&. The 

literature on the dating of this poem has already become enormous. Hu Shih tends 
to date it earlier than most others, about the middle of the third century A, D., cf. 
SLftiPffo, 6:149.9-14: ?Li£^^?fe#J^f£. Others tend to date it much 
later, about the fourth and fifth centuries A. D. 

87 E. g., among the Mi wok, where marriage with the wife's brother's daughter is 
reflected in twelve terms [E. W. Gifford, Miwok Moieties, University of California 
Publications in American Archaeology and Ethnology, 1916, p. 186], but these are all 
lacking in the Chinese system. 

98 Among the Omaha, marriage with the wife's brother's daughter is reflected in 
the conceptual identification of the father's sister, the female ego, and the brother's 
daughter [A. Lesser, Kinship Origins in the Light of Some Distributions, American 
Anthropologist 31 (1929), 711-12], but it is not indicated in the terminology. In the 
Chinese system the generations of father's sister, husband's father's sister, and the 
husband's sister are clearly distinguished conceptually, although the terminology fails to 
differentiate them. 

99 Tso chuan jR-fiCH-H^p, 35. 18a. 


ego marries wife's sisters. Ego again marries father's widows aft^r 
his decease. Then mother's sisters become equated with wife's 
sisters. This explanation seems fantastic but has some support, 
since a man's secondary wives (concubines) may also be called i, 
i. e., mother's sisters, wife's sisters, and secondary wives are all 
grouped in one class. It is well known that the sororate was prac- 
ticed among the feudal nobility, but, as regards the inheritance 
of father's widows, there is no authenticated evidence. 100 Indeed, 
such a marriage would have been abhorrent to the ancient Chinese. 
We learn that the old writers detested the Hsiung-nu, pastoral 
nomads of the northern steppes, who married their fathers' widows, 
and never failed to mention this as an excuse for derision. 101 

K'ung Ying-ta [A. D. 574-648] explained the extension of i to 
include mother's sister (ts'ung mu) as due to the psychological 
similarity between these relatives. 102 Mother's sister's are i to 
one's father, just as wife's sisters are i to oneself. The son, imitat- 
ing the language of his father, also applies i to his father's *. In 
short, this case seems to demand a psychological explanation, 
together with reverse teknonymy. 

The connotations of shin are likewise of significance. Father's 

100 M. Granet cites the case of Duke Hsiian [718-700 B.C.] of Wei who married his 
father's concubine I Chiang (La civilisation chinoise, p. 401). Apparently, Granet is 
not aware of the fact that Ku Tung-kao jgg^jltj [1679-1757] has convincingly shown 
that I Chiang had not previously been Duke Hsiian 's father's concubine (Ch'un ch'iu 
ta skih piao ^ft;M^, tfiti&M ed., 50. 3a-4b: W^^W^IcP) ■ Even if 
it be admitted that this is an authentic instance of step-mother and step-son marriage, 
nothing is proved thereby, since the instance is quite anomalous. Not only are such 
anomalies recorded in quite a few instances from the Ch'un Ch'iu period, but also 
examples of incest, involving actual blood relationship, like grandmother and grandson, 
brother and sister, cf. BEJTJfl* SrS&HlS- But Granet has omitted these. Indeed, 
such anomalies are always cropping up in Chinese history, even in quite recent times, 
e. g., the T'ang Emperor Kao-tsung [650-684] married his father's concubine Wu 
Chao jr^H later known as the notorious Empress Wu of the T'ang dynasty, and the 
Emperor Hsiian-tsung [712-756] married his son's concubine Yang Yu-huan, well- 
known to Europeans as the foremost beauty of China. One wonders how Granet would 
interpret these instances. 

101 Shih Chi mSLMW. "0.2a: £###&#, ft fcJEfMfc££££. 

The Hsiung-Nu customs of marrying father's widows other than one's own mother and 
the levirate were so well known to the ancient Chinese that they frequently men- 
tioned them as a sign of the moral inferiority of the Hsiung-nu. 
* Tso Chuan JI^ZH-H^p, 35. 18a. 

102 I 


younger brother's wife and husband's younger brother's wife are 
both called shen in the vocative. In the Erh Ya father's younger 
brother's wife is called shu raw, which is the standard term today. 
Shen first came into use during the Sung period; it is usually 
regarded as a contracted pronunciation of shih mu. 10z Its exten- 
sion to include husband's younger brother's wife was effected 
about the same period. 104 No possible marital relationships could 
give rise to such an equation of relatives, nor could any other 
sociological factor. Teknonymy offers the simplest solution. Hus- 
band's younger brother's wife is the shen of female ego's children. 

The origin of the terms kung, for husband's father, and p'o, for 
husband's mother, has never been investigated. The old term for 
husband's father is chiu, and for husband's mother, ku> both of 
which reflect cross-cousin marriage. During the first millennium 
A. D. a large number of terms were introduced for the designation 
of these two relatives, since chiu and ku were no longer applicable 
after the discontinuance of that particular type of marriage, but 
kung and po finally gained prevalence. 105 

On the other hand, kung and p'o are prevalent grandparent 
terms. 106 Why the wife should apply grandparent terms to the 
husband's parents is rather perplexing. This terminology is most 
susceptible to marital irregularities, but we cannot see what mari- 
tal form, no matter how startling, could be involved here. But if 
we assume teknonymy, the situation immediately explains itself. 

The foregoing cases comprise the most significant terminological 
anomalies, and constitute about all the instances in the Chinese 
system, both ancient and modern, where the generation principle 
is openly violated. In every case we have tried to explain these 
exceptions by facts and hypotheses which have proved illuminat- 

"*Ming too tsa ehih, 13b: £M£W,jRtt&**#£ ■ . - #^fft,*§^75 
"UttS^ .^tt P?Hfi» • Shih mu, pronounced as one word, became shim, or shen. 

104 Tzu wei tsa chi %frM$ttl> of Li) Tsu-ch'ien gMfl> 1187-1181 A. D. [Shuo fu 
tftft. 19] 2a: g&**S, $&$«gJ§tt?F, #«*»***. 1$%^® 

^^I^Pf M y #lf ^F^H^ jWfrl^I ♦ Mu mu is used in the sense of husband's older 
brothers wife, and shen in the sense of husband's younger brother's wife. 

"■ See Table IV, terms 1 and 2. 

109 See Table I, terms 13 and 14. 


ing in the discussion of analogous phenomena elsewhere. But we 
have found none of them applicable to the Chinese situation; 
rather, we have found teknonymy the only satisfactory explana- 

There can be no doubt that teknonymy is the determining fac- 
tor in all these cases, but one may ask whether teknonymy is 
universal in China and of sufficient antiquity to have been in- 
volved in producing such effects in kinship terminology. The 
universality of this practice in China is unquestionable ; the fre- 
quency of its use, however, might have varied in time and place. 
At times the practice was so accentuated that the ordinary forms 
of address became hardly intelligible. 107 In many regions, e.g., 
Wusih, Kiangsu, the bride ordinarily addresses her husband's 
relatives as if she were one generation lower. The teknonymy 
practiced is usually of the type that omits the child's name, just 
as in English a man may call his wife simply " mother." This 
type is especially efficacious in producing the irregularities just 

As regards the antiquity of teknonymy, we must depend upon 
historical evidence. The earliest instance that can be interpreted 
as teknonymy is recorded in Kung-yang's commentary on the 
Spring and Autumn Annals [Ch'un Ch'iu]. In the sixth year of 
Duke Ai [B. C. 489] is recorded an instance concerning Ch'en Ch'i. 
He, in referring to his wife, says, " The mother of Ch'ang. . . " 108 
Ch'ang is known to have been Ch'en ChYs son. The teknonymous 
usage here is indubitably clear. This instance, in the fifth century 
B. C, is more than a millennium earlier than most of the cases we 
have discussed above, except those involving i and sheng, which 

107 E.g., the extreme instance of the practice of teknonymy in South China, as 
recorded in the Ch'ing hsiang tsa chi fl^Hfil, by Wu Ch ( u-hou i^J&Jli ca. 1080 
A. D. [fi^fit edj S.2b-Sa, of the Sung period, which reads: Sfgja/fS^P^ 

*f Bf*^H [modern gfi|fi in Kwangsi] £&, *f ftlfeg, J§£-£, 6PP?M 
^Kung-yang chvun &&^*?, 27. 12a: ffcfc^ffiSg. Rfci! B : liKMfc 


are about contemporaneous. On the other hand, if we make 
allowance for the conservative spirit of the classical writers in 
recording colloquial language, it is reasonable to infer that tek- 
nonymy is much older than this documentary evidence would 

The influence of teknonymy on kinship terminology is quite 
apparent. Gifford, in discussing similar usages in the English and 
Californian kinship terminologies, has cogently remarked: " There 
must be, in other kinship systems, many analogous cases [cases 
analogous to the English teknonymy of calling husband " daddy " 
and wife " mother "], some of them crystallized into invariable 
custom like the Luiseno case, cases which require no startling form 
of marriage for their explanation, but which could be readily 
understood as our own, if we were but familiar with the family 
of the group in question." 109 

In the very limited literature on teknonymy, various theories 
have been put forth to account for its origin, but no serious 
attempt has been made to use it in explanation of other social 
phenomena. Teknonymy as a usage is based on kinship and 
kinship nomenclature — a circumlocutory way of expressing em- 
barrassing relationships. Through long and intensive use, why 
should it not have produced certain peculiarities in kinship termi- 
nologies, as other social usages are reputed to have done? The 
Chinese cases are especially illuminating. It would require a 
series of marital or other special practices to explain the pecu- 
liarities of chiu, po, shu, ku, i, shen, hung, p'o and sheng, whereas 
they can uniformly be explained by the single principle of 

109 Californian Kinship Terminologies 265. In the Chinese instances the situation 
is so apparent that even so amateur an observer as Hart has been led to remark: *' The 
nomenclature employed in the designation of two brothers-in-law and two sisters-in- 
law, i. e., by a wife toward the brothers and sisters of her husband, and by a husband 
toward the brothers and sisters of his wife, seems to have its origin in the names 
applied to such people by the children (their class children, or nephews and nieces) 
born of the marriage. Thus an individual's wife's brother is the kew [= chiu] of that 
individual's children, and that individual in speaking of him as his brother-in-law, 
employs the same word, kew, to designate him as such, so with the other." [Morgan, 
Systems . . . , p. 413.] Morgan, however, engrossed in his evolutionary " stage 
building," entirely overlooked this pertinent remark. 


The Chinese method of counting relationships starts with the 
three nearest kin, i. e., parent-child, husband-wife, and brother- 
sister, and extends out in all directions. Whenever one comes to 
the question of Chinese kinship extensions, one always stumbles 
upon the problem of chiu tsu AM, "nine grades of kindred." 
This is a much discussed but vague term, which first occurs in the 
Book of History, 1 Its interpretation comprises two major theories, 
representing two different schools of classical commentators. The 
Modern Script School interprets the chiu tsu as follows: The 
four groups of relatives of the father; plus the three groups of 
relatives of the mother; plus the two groups of relatives of the 

The four groups of relatives of the father are: 1. With ego in 
the center, counting four generations above, four generations 
below, and the four collateral lines each counting four generations 
from the lineal line from males through males. 2. Father's sisters, 
when married, and their descendants. 3. Ego's sisters, when 
married, and their descendants. 4. Ego's daughters, when mar- 
ried, and their descendants. 

The three groups of relatives of the mother are: 5. Mother's 
father and mother. 6. Mother's brothers. 7. Mother's sisters. 

The two groups of relatives of the wife are: 8. Wife's father. 
9. Wife's mother. 

This interpretation is not followed by the Ancient Script School 
of classical commentators, who believe the chiu tsu includes only 
the sib relatives but not the non-sib relatives. Therefore, accord- 
ing to their interpretation, the chiu tsu takes into account only 
the first of the above nine groups of relatives. That is, the chiu 

1 Shu Ching f§-H, 2. 7fc: J^l^A^- Whether this is the earliest use of the term 
or not is very questionable, since the antiquity and authenticity of the §§:P!- is much 
questioned. However this is the focal point of later controversies. HpJtfil] has a 
very penetrating discussion of the Problem of Chiu tsu j/L$^PfJIB» in ^^[^©"Pj 
37: 9-10, 105-111. 


tsu means simply nine " generations," viz., the four ascending and 
the four descending generations, with ego in the middle. Naturally, 
because of the collective responsibility of the individual's social 
actions, many students tend to narrow down the interpretation in 
order to lessen the social and legal complications. The chiu tsu 
problem, however, is purely an academic and historical matter. 

In the present work the classification of the Erh Ya is generally 
followed. The actual extent of the terminology listed, however, 
is entirely dictated by the needs of modern research and by the 
material available, although all those relatives within the Chinese 
mourning grades are included. The sib relatives are emphasized, 
but such emphasis is unavoidable in a system based strictly upon 
a patrilineal social grouping, where many of the terms for non-sib 
relatives are merely extended combinations from those for sib 

The relatives are divided into two main groups and subdivided 
into four tables. The two main divisions are Consanguineal Rela- 
tives and Affinal Relatives. Under Consanguineal, there is a sub- 
division into Relatives through Father and Relatives through 
Mother ; and under Affinal, into Relatives through Wife and Rela- 
tives through Husband. Under each group are also listed persons 
connected through marriage, that is the " in-law's " ; this is a 
conventional practice. 

In the tables the modern standard terminology is given first 
under each entry; these terms represent the system as it now 
stands. The slight variant combinations in some of the com- 
pounded terms have been carefully collated against one another 
from contemporary sources. In collating, two criteria were 
adopted: the statistical and inferential; i.e., if two or more forms 
are equally common, that form inferentially most in keeping with 
the working principles of the system is given as the standard 
form, and the others are given as alternatives. These variations 
are rather insignificant, and, in view of the number of people 
using the system and the geographical extent of the system, 
quite inevitable. 

Under each standard term are given the historical terms in 
their chronological order. Their exact nature, whether alternative, 


literary, or dialectical 2 is indicated, together with their modern 
status. 3 Following them come the referential and vocative termi- 
nologies. If no terms are given, it is understood that they can be 
formed from previously stated formulae. No stereotyped order 
is followed, but generally the treatment varies according to the 
nature of the material and in keeping with special circumstances. 
Citations illustrating the use of terms are given in the notes. 
It is possible only to give citations for the earliest occurrence of 
a new term, or for an old term used in a new meaning. In some 
cases, the most typical instances are cited. These citations are of 
importance, since only through them can the exact nature and 
chronology of a term be determined. 

In the tables I have attempted to record as fully as possible the 
whole range of terms. 4 Thus, under the father's father no less 
than twenty terms are given, under various usages and periods, 
and in some instances the number is still greater. Of course, not 
all terms can be so treated, especially terms for distant relatives, 
which are compounded from the basic terminologies, but the slight 
variations in any possible combinations are given. A full record- 
ing of the whole nomenclature and the determination of the exact 
nature of each term are indispensable conditions for the proper 
understanding and interpretation of the system, since most of 
the early misunderstandings are the result of partial and mixed 
renderings of the system comprising forms from different strata 
of the terminology. 

2 Dialectical differences in Chinese kinship terminology have been somewhat ex- 
aggerated. Many of them are mere local variations in pronunciation which do not 
affect the morphology of the system. 

3 1. e. f the exact connotation and nature of the term in modern usage, if it is 
still used. 

4 Terms that are idiosyncracies and have no general currency are excluded. E. g., 
the Emperor Hsiian of the Northern Chou dynasty did not like others to use the 
term kao, so he changed the term kao tsu to chang tsu, and tseng tsu to tzu chang 
Uu:Pei,hih JHMjB, 10.32a: X^SSgA^ &:#- *^. • • .3fc • • • % 
^^rfiM^^Jfcilfi. H"lfi3$3KJIilL Terms of such nature will not be considered 



Relatives Through Father 
I. Generation of the father's father's father's father 

1. Kaotsufu m&Xiiii 

In the Classics kao tsu was sometimes used to mean any ancestor 
ascending from grand-father. 1 In the 17th year of Duke Chao 
[5%5 B. C] in the Tso Chuan, the first ancestor was called kao 
tsu. 2 In another place the ninth ancestor was referred to as kao 
tsu* King K'ang [1078-1053 B.C.] called King Wen and King 
Wu kao tsu, but actually they were his great-grand- and grand- 
fathers. 4 Apparently, during the Chou period, if the actual father's 
father's father's father is meant, wang ju must be added, like the 
kao tsu wang ju used in the Erh Ya. The term kao tsu does not 
occur in the mourning relations of the J Li 5 and it is surmised 
that any lineal relative ascending from tseng tsu may be called 
tseng tsu? 

During the T'ang period kao men 7 was used, but infrequently ; 
it was most likely a posthumous term. The posthumous and 

*m cMhiu, 24.ia: &ffiJa#jl;£££iiSJilL #;£#MS> iSJil* Site 

* Tso chuan 8^+4;^, 48. 5a: ^0 - - . #it&JH*M£ *?£:£*» ■ • • 

*ibid„ Bg£-h5^, 47.11b: 3EB . . . a#flffi«Ja«ffilR^W^*». 

fiUSrfciifc, $C0H K- The term iUlM is meant for the ninth ancestor of f£fife 
It is also used in the same way in the Shu Ching, e.g., $S J| 9. 17a: ^Jt^^^ 
ffti&M^lffi' The term ^° tsu refersto Ch'eng Tang (1765-1760 B.C.?). 

'Shu Ching, Jg3E;£!§, 19.3a: ife$|§lM- 

6 In the |J|Jjg/f|| of the / Li, since there are no mourning specification for f f f f, the 
term for him does not occur. In the ritual works, the first use of kao tsu for f f f f is 
in the Hjflft/hte of the Li Chi, 32.7b: ^^HtM^'Z^ ftSHSSfi^-lfc- 

•Ming ch'i pi fan, 3.3a: ||JlM#i&#^, M38§r&l!£«- «% M- 

gMM-b ^#ffl-&; a«MT.e#«*;SI"ffiB:»Hb. Very probably 
the use of kao tsu for f f f f , and tseng tsu for f f f is due to the development of the 
sib organization of the Chou period but originally both of them meant simply " distant 
ancestors." Cf . T Ul I ^£felK#. CYYY, 4: 4 (1934), 399-415, in which he con- 
siders that during the Shang and early Chou periods, exact relationships are only 
counted to two generations both above and below. 

'Chin shih ts'ui pienJ&ffMW* 101. lib: . . . jgP , J3*lgj£ , Jt:£#5fe, 
Ch'eng wei lu, 1.4a, S?l©P , !» igjUWl- 


temple term used in the Li Chi is hsien kao? The term is no 
longer used in this sense. Since the Yuan dynasty (1280-1367) 
hsien kao had been used as an epitaphic term for father. 

2. Kaotsumu iftffl-Brfffm 

Kao tsu wang mu iSMI^ is the term used in the Erh Ya. 

3. Kao tsu ku mu ifcjftifrl* f f f f si 

Kao tsu wang ku rftffl.BEiS in the Erh Ya. 

4. Kao tsu ku fu ift »**£ f f f f si h 

No term is given in the Erh Ya. In tracing relationship through 
women of the father's sib married out, the terms for these women 
are usually employed; the terms for their husbands, only infre- 

II. Generation of the father's father's father 

5. Tseng tsu fu #i&3£ fff 

As a form tseng tsu may be used alone. Tseng tsu wang fu is 
the term used in the Erh Yd. Tseng ta fu, 9 ta wang fu 10 and 
wang ta fu u have been common alternative terms since the sixth 
century A. D. Tseng men™ was commonly used during the T'ang 

8 Li Chi «H*,46.8M>: fi|#J$f. $: EHgtJfj^ Aft-tiL- Oft. «*§, 
9. 18*13: ££«£. «&: ft ftftfe. JkMPMMBM, SftjSi 

>shih chiMtt, «.ib: S££tt, »££#»& gjJpf£3£B a« • • . 

S^#^Kg«AO«, -g^^ffiflffi. CI'*** cW&fp^gfl}, 24.1b: 
H^^^DiH- • • • [Here and in two following notes supply No. 4, on p. 149.] 
10 Chu-chiang wen chi ^jlll' 19* 3b: jtlLOjik- • . ■ 
"Chin shih yao li #!&0#J, 7b: J^g^fflW JfiS^M^I*^ 
™H S in Tang shu #£gfteM|, 195. 6b: . . . i5fc#ifflja#, ^+— ^ 
75*. CAm t*t& fa'w P ien JkSRJ&MMUIH. 82.13a-19b: OTS^HB^lfr 
. . . Chien Ta-hsin [Chien yen t'ang chin shih wen po \wei, 6. 13] says: fflitMLlEb 
hH' ^SfStSS- From what can be deduced from the evidence, men f^ was a 
very common posthumous term in referring to lineal relatives from the second ascend- 
ing generation and upward during the fourth to the eighth centuries A. D. Father's 
father was called sfcPJif*, father's father's father's father was called fftP^, so father's 
father's father was called tt1"J . Its origin may be similar to the term t£^t!fe 
PJ + ^ explained by Yen Chih-t'ui. See p. 222, note 132. 


period. All these terms may be used posthumously. The ancient 
posthumous term is huang kao, lz as used in the Li Chi. Huang 
kao was later used as a posthumous term for father, but has been 
prohibited since the Yuan dynasty, since huang implies " imperial." 
The vocative is rather variable. Tai wing 14 and tseng weng 1S 
were common from the fourth to the ninth century A. D. Weng 
means " venerable old man." Tai kung and t'ai yeh yeh 18 are 
common modern vocatives. 

6. Tsengtsumu#M#ffm 

In the Erh Ya, tseng tsu wang mu. The most common modern 
vocative is t'ai po, or t'ai po p'o. 17 

7. Tseng po tsu fu fiBȣ f f f o b 

Tsu tseng wang fu is used in the Erh Ya, and tsu tseng tsu fu 
in the / Li. 18 Both these terms also apply to father's father's 
father's younger brothers. 

8. Tseng po tsu mu 1H6 11© ff fob w 

In the Erh Ya, tsu tseng wang mu, and correspondingly tsu 
tseng tsu mu in the I Li, Both these terms also apply to father's 
father's father's younger brother's wife. 

9. Tseng shu tsu fu #&ȣ f f f y b 

The vocatives of 7 and 9 are similar to 5, but differentiated by 
prefixing their numerical order or titles. 

"LiCAi^$~,46.8a-9: B«Ji ?Lgfc- 0:g#ffi* #*&*• 

"NanShih mB%T^WM.&, SI* ±m* 

15 ^tm- Since weng was commonly used as vocative for grandfather, tseng weng 
~.nd t'ai weng were used for great grandfather. Weng may also be used for any 
venerable old man. 

16 jfc& an d >fcj§ J& ■ T'ai means " great." Kung and yeh yeh are vocatives for 
grandfather; t'ai kung and t f ai yeh yeh f for great grandfather. 

" jfc^ic °r >fc§l§?- P'° or P f ° V° are vocatives for grandmother; t*a% p*o, for 
great grandmother. K'ung Fing-chung JL^W. ca. 1080 A. D., ^tJcJfe, MM 

18 1 Li Mb* 33.6a: ^#fflO#- »&:]&#«£#* Mgjfc£« 



10. Tseng shu tsu mu #&&# f f f y b w 

The vocatives of 8 and 10 are similar to 6, but differentiated by 
prefixing their sibnames or their husband's numerical order. 

11. Tseng tsu ku mu MJfrtt f f f si 

Tseng tsu wang ku is the term used in the Erh Ya. Modern 
vocative, ku t'ai p'o. 

12. Tseng tsu ku fu #«**£ f f f si h 

Modern vocative, ku t'ai kung $&$:& or ku t'ai yeh i^isM. 

III. Generation of the father's father 

13. Tsu fu MX f f 

Tsu may be used alone to mean father's father, but it may 
mean any ancestor. 19 In the Erh Ya, tsu was used synonymously 
with wang fu, 20 but wang fu is now more often used in the 
posthumous sense. Since the Han period ta fu 21 has been fre- 
quently used. Other early alternative terms, like tsu chun 22 tsu 
wang fu 23 and tsu weng 2 * are commonly met in literature. An- 
other common early term is kung 25 which in many localities is 
still used as a vocative. 26 Since kung is a common complimentary 
term for any older man, its connotation as a relationship term is 
indefinite. It is used to mean father, husband's father, etc. T'ai 
kung 27 is used in the Hou Han Shu to mean father's father, but 
is now used as a vocative for father's father's father. 

» J* chih lu, 24.1a: g ^ffif -b ^HM- »|RiP2^0, ^jfififtM*,- 

30 Erh Ya:M> lO-lfc. 

^ShihChi 8fe1k&.5B.laz «««&* ##$»A*r *£IH«1- ■ - . 

ibid. ftgrpftf, 120.6b: mmn3t%MizOtf ■ 

"Kung ts'ung tzu ^fl, 2.45b: ^®g££, B . . MSM**Bl#ft£5£ 


23 Chin shih ts'ui pien ^.X^MU^ 113.02a: ^£IMiS87CflE3!F£§*& 

34 ^^m^u^mw^m^Mmj&^'zm- <& c^ *« *«, 1. ? a . 

25 Lu ,*& C V«« cAVu, 10. 10a: ft?£*iPffc$£;fr#:#, ft^ffctffif PJ2 


28 Cf . Ch'eng wei Lu, 1 . 8a. 

"ffou flan SAu «£@{|, 93.14b: $ifr£i|5 |[@^ . . . %-%$§ ■ ■ . 0: 


The most common modern vocatives are yeh yeh MM , kung 
kung &&, a weng Wlm, wing weng mm. 2S Their usage depends 
upon local custom. 

The depreciatory term is chia tsu or, uncommonly, chia kung. 
The prefixing of chia to tsu to refer to one's own father's father 
dates from the Han period. It is sometimes condemned as in- 
correct and vulgar, 29 but is nevertheless in universal usage today. 

The complimentary term is tsun tsu ju?° Ta men chung 31 
was used as a complimentary term during the fifth or sixth cen- 
tury A. D. It was most probably a posthumous term, and is not 
used today. 

The ancient posthumous terms were wang kao 32 and huang tsu 
kao. 33 The former was probably more often used in connection 
with ancestral temples, and the latter, with sacrifices. This minute 
distinction might be due to the important role of ancestral wor- 
ship and sacrifices in the sib organization during the feudal period. 
During the fifth century A. D. hsien wang chang jen 34 was com- 
monly used. None of these terms is used today. The modern 
posthumous terms are hsien tsu ^fejffl. or wang tsu "CM. 

14. Tsumultfm 

The term in the Erh Ya is wang mu 3M$, which corresponds to 
wang fu. Later terms, like ta mu 35 and tsu p'o™ more or less 

^RJ** 1 - g±M*, MMt, ffi^&jfc? During the Han period, 
kung was sometimes used for father; t'ai kung, for father's father. 

2S Shih shuo hsin yii, SB. 10a. 

" Cf. Yen shih chia hsiin JE^Itt. 2- 4b-5b. 

'"Ibid., 2.5a-b: JLHAW- fiMBfiOflfc U£0#, ©#££#, £*D*^. 
g &OT> JM*DJt#!. #^^H-ul. [Supply No. 4, p. 149.] 

n Ibid., 2. 6a: ^"J tp . 

"Li Chi &&, 46.8a-9b: 0£#ij. lift: 0I#I* JftftHb- 

"Ibid. ftflf ,3.22a: ^IOHII#, 3E#0£J8L*lt 

" Yen shih chia bun #*£!£, 6. 15b: ^fltfSP^M^&trjfcA. %&± 

"Hsin shu ft%$, 3.1b: *#«#, £lj*©£> ftS*^£- Han shu £= 


80 !&§?; see p. 210, note 24. 


correspond to the terms for father's father. The depreciatory 
term is chia tsu mu, and the complimentary term, tsun tsu mu. 
The ancient posthumous term is huang tsu pi, 31 and the modern 
term, hsien tsu mu. Tax po was once used as a vocative for 
father's mother but it is now used for father's father's mother. 
The most common modern vocatives are p'o po §<§! 38 and nai nai 


15. Po tsu fu f6M3£ ffob 

Both the Erh Ya and I Li give the term tsung tsu tsu fu. 40 
It applies to both older and younger brothers of father's father. 

Po weng 4t is a modern vocative. The vocatives of 13 may also 
be applied here by prefixing the numerical order or title. 

16. Po tsu mu ibWt f f o b w 

Ts'ung tsu tsu mu is the term both in the Erh Ya and the I Li. 42 
Since the Erh Ya also gives the term ts'ung tsu wang mu #£!!:£#> 
in can be inferred that during the Chou period father's father's 
brothers may also have been called ts'ung tsu wang fu. During the 
Han period it was abbreviated to tsung tsu mu. 43 Po p'o 44 is a 
more modern term, and may be used vocatively. 

17. Shutsufu^ffl^ ffyb 

Compare 15. In the Kuo Yu the term ts'ung tsu shu mu 45 is 
used for father's father's younger brother's wife. It can be in- 
ferred that in ancient times tsung tsu shu fu may also have been 

37 Li Chi ft ff, 3.22a: I#0J|0#t 

-K'ung Fing-chung $$&, op. cit, ftjbtt&lf$l$t. 2.19a: $$£| 

"Ck'in ,hu chi 1.6b: ft, tifrWMB®' H**- i£#«> *J3W O 
"PJo i^fy^ffi^fi^- M 1S & l s0 written jjj). It originally meant mother and was 
read ni, as in the Kuang yiin, My SLAP^J^- 

*° I Li M$> 33- la: fl£0 OOf 

41 f6ia> cf > H^S* Jt^Hfil in Pa ch'iung shih chin shih pu cheng, 117. 12a- 

48 Cf . note 40. 

«LiCW«3. 9.85b: #£B: »A*fft- M^ «£* MT'&O . 

"Kuo yu %Wr> 5.12a: ®tO%i&~Z^ ^^fltO&M . [First miss- 
ing character is No. 4, p. 149.] 


used for father's father's younger brother, and ts'ung tsu shih fu 
for father's father's older brother, but this inference is by no 
means certain. Since the Tang period shu wing 46 has been a 
common alternative term, and may be used in the vocative. The 
vocatives of 13 may also be applied here, according to local usage, 
by prefixing the numerical order. 

18. Shu tsu mu MEM f f y b w 

Compare 16 and 17. In the Han dynasties chi tsu mu 47 seems 
to have been used. According to another interpretation, chi tsu 
mu means father's father's brother's secondary wives, 48 but there 
is no means of checking this. Since the Tang period shu po seems 
to have been a common vocative. P'o p'o may be used in the 
vocative for 16 and 18 by prefixing the numerical order of their 
husbands, or their own sibnames. 49 

19. KutsumuiSM* ffsi 

Wang Jcu 3E#i!f is the term given in the Erh Ya. 

20. Kutsufu*£M3£ ffsih 

SI. Chiu tsu fu JtM3t f mb 

Chiu tsu 50 may be used alone. The inverted form tsu chiu 51 
was used quite early; and ta chiu 52 was used during the Later 
Han period. 

22. Chiu tsu mu 00# f m b w 

**Ch<ang-li chi. 23.13a: tg^fc^+Mfi^JC' Han Yii refers to himself as 
+AMm and t0 his wife ^ +A&SS- Ibid., 23. 12b: %%$!%, where he refers 
to himself as ~\~ /VHj and to his wife as ~h/\J§c- They were offering sacrifice to 
his brother's grandchildren. 

47 Chin shih tsui pien 18.1b: lKt>ff^O#. 

"Ch'ien Ta-hsin says [Ctiien yen tang chin shih wen po wei, 1.26b] that j£ff§ 
^0#, ^SWJ^O#4fc. [Supply tsu, ancestor.] 

" Yen shih chia hsiin jRJftJS, 2.7b: 0#£-tfc&#, %#to3&tiE&L$i2.- 

80 Leng In tsa shih J^Jg^g^, by Lu I-t'ien Ht#?IS. ca. 1850 A. D. UfS/J> 

SfcMR ed. 2.25b: -fcfcfRffi • ■ ■ ^lO^O^O . [Supply fu . . -. chiu . . . 

chiu-tsu 1 

n chin shu MM, 70. lb: 8tK!;fcJfMMfil3;. M±o O-tiL, 

"Hon Han Shu gftftft. 74.2a: 00^, Mi$MM]M.%*:A ...ft... 

£,#£, ftStkWH: ^Mkfck®^. [Supply uu-M 


23. I tsu mu ^ffl# f m si 

24. I tsu fu 003: f m si h 

The vocatives for the above can be constructed from the voca- 
tives for grandparents, whatever forms are used, by prefixing ku, 
chiu and i. 

25. Tang po tsu fu ^f6M3cl s of 7 or 9 > f f 

26. Pang po tsu mu OOOtt w of 25 

27. Tang shu tsu fu SftftX s of 7 or 9 > f f 

28. Tang shu tsu mu OOO* w of 27 

No term is given for 25 and 27 in the Erh Ya, but tsu tsu wang 
mu ^ffll'ft is used for 26 and 28. It is inferred that tsu tsu wang 
fu may have been used for 25 and 27. In the / Li, tsu tsu fu is 
used for 25 and 27, and tsu tsu mu for 26 and 28, 53 whereas in the 
Erh Ya these terms are used for the son and the son's wife of 
either 25 or 27. The I Li is simply using an abbreviated form. 

Vocatives of 15-18 may be applied here, respectively. 

29. Tang ku tsu mu £J£iL» d of 7 or 9 

In the Erh Ya, tsu tsu ku. This term is rather inconsistent with 
the whole Erh Ya system, since logically it should be tsu tsu 
wang ku. Perhaps the Erh Ya was already beginning to use 
abbreviated forms. 

30. Tang ku tsu fu ^ii^ILX h of 29 

Vocatives can be built up by prefixing grandparent vocatives 
with ku, and can further be differentiated by prefixing the numeri- 
cal order of 29, or the sibname of 30. 

31. Piao tsu fu*ffi£fffsis 

32. Piao tsu mu 00# w of 31 

The above two terms can be further differentiated by prefixing 
ku, chiu and i, e. g., ku piao tsu fu for f f f si s, chiu piao tsu fu 
for f f m b s, i piao tsu fu for f f m si s. Ordinarily, piao tsu fu is 

"/ Li, 33.6a. 



applied to them all. Usually these relationships are not continued 
socially after the death of the f f f si and f f m, unless it be that 
either of the parties strongly wishes to maintain them. Terms 
for their descendants will not be given in this table. If the relation- 
ships are maintained, terms could easily be constructed, e.g., the 
sons of 31 will be called fang piao po fu and fang piao shu fu, etc. 

IV. Generation of the father 

33. Fu % father 

Fu is primarily a standard literary term throughout, and is 
seldom used alone as a vocative. Weng 54 was an old vocative. 
A kung 55 and tsun 56 were prevalent from the third to the sixth 
centuries A. D. About the fifth century A. D. members of some 
of the royal families called father hsiung hsiung, 57 a term for 
older brother. During the T'ang dynasty the royal family called 
father ko, 58 nowadays a universal vocative for older brother. Be- 
fore this, ko had never been used in either of these senses. 50 It 
might be that ko was an old dialectical term for father, and that 
during this period it became confused with hsiung, thereupon 
losing its original meaning of father and acquiring the connotation 
of " older brother," but the matter is most perplexing. 

84 Shih chi m%l*fc> 7.26a: g|IS . . . ^6P5g^. 

"Nan Shih MW2i^> 34.4b: . . . X^SI^r^- According to ^&M 
[Kuang ya shu cheng] $* and ^j are very similar in sound and may be dialectical 
renderings of a same term. 

"Sung ,hu Mt®^<$, 67.30a: M^^^iaUt' rfffW^M^ Shih 
shuo hsin yii ggff, 2B. 30a: flJfJ-Mi^&fHSt, Ut^^^+S, ftft 

*£g§. m&, w\qb, fij?*i£fpi*o#? ibid, ob: m±w%.m< ^mmtn^. 

^&7i?W> &r\,mm^U:fa%®im& [Su PP iy/«.] 

"PeiCh'ithu mi%3.ffiM. 12.6b: Rfc 

"Chiu Tang shu 5^#, 106.16b: 7C5rSt&0 : Hlftl^. The ko is meant 
for the Emperor #Jj£. Ibid. J$3i$fl|, 107. Sa: 'ft— If^^H- The -^ 
is here used by S$ in referring to his father, Emperor Hsiian-tsung. Ko was also 
used by the royal family of the T'ang dynasty to refer to oneself before the son. 
Ch'eng \wei lu, 1. 15a: states: ^MWijf ±£HiS£*liSlf VtM- © #T^ 
SffiWW, HMffe***lfi- Supply fu.] 

59 ?f is defined in the Shuo wen as M to sing," or M a song." 



Yeh M is a vocative used from the sixth century A. D. on 
it is also written M. 61 Tieh W* 2 which may be a later variant 
pronunciation of t'o i£, 63 a dialectical form of western Hupeh of 
about the same period, is now a common vocative. Pa pa is al- 
most as commonly used as tieh ; it first occurred in the Kuang Ya, 
which reads, " pa ... is father." 64 The Cheng tzu t'ung considers 
pa a term of the southern aborigines, 65 and states that the ab- 
origines call their elders pa pa AA, or pa pa BEL, and that the 
Chinese lexicographers added the classifier 3£ to form l£. On 
the other hand, pa is also considered a later variant pronunciation 

The dialectical difference, insofar as the evidence goes, seems 
to indicate that tieh is predominantly a northern usage, 67 and pa, 
a southern ; 68 but this explanation is by no means certain. 

The Kuang Yun states that the people of Wu call father che* 9 
but according to the Tung Ya father was called lao hsiang in 
Wu. 70 None of these terms seem to be in use today. The in- 

90 Yen shih chia hsiin %j^|f , 4. 10b. 

"Nan shih 4^41, 80.22b: 3Eft® il>t;Ji • ■ ■ 3£ff^« ■ • ■ &H: 

BUtfrlR^tgti, '&WM%W:- The original form is JflJ, and jg is a later form 
with classifier No. 88 added. For discussion cf. Kai yu ts'ung k'ao, 37. 15; Heng yen lu, 
3.2; and Ch'eng rwei lu, 1.15-16. 

"Ch'ang-U chi &-fc^-£r£, 23.14a: H^HA. ffi&ifr ■ • • ^jglS* 
Wfr-ZM- Cf. Shu p'o, 1.2b-3a, n®'$M- 

"Kuang ya: £ . . . 0-&, pronounced t'o. Nan shih ^n^,fe^HSiS. 52. 

7Kik : fflB*ffi3fc"*?LJ»- flJ±*sIiO^^, iftS- For discussion cf. Kai 
yu ts'ung k s ao> 37. 15. 

"S . . . 04 t 

85 Cheng tzu t'ung: ^ . The Miao and Yao tribes of southwest China still call father 
pa, or its slight variants. Cf. Miao fang peilan ^Kfliift; by Yen Ju-i |g #U!)3?[ 1843 
%n^t1%? ed -J 8 - 6a > 9 - 2b > anc * 9.10a; Ling piao chi man ^^fil® by Liu Hsi-fan 
$\W)M [1932, Shanghai, Commercial Press] 137. 

66 Cheng Chen (Ch'in shu chi, 1.1b) says that T^lfB^qii' t[JO ^Itffif > 

e8 C/*i Fun: f£, (^«r-tJJ, X^H-tJJ) ^A^O. The CA'en ff wei lu, 1.27a: 
also states that j^^^O ^p^ffi- ^ fS ma y be a different rendering of pa |§. 
flB ft (JE^-BJ)? ^AP^O • The modern pronunciation is the same of yeh f^. 
70 Tung ya fjglg, 19.4b. 


habitants of Fuchow call father lang pa fitfH ; 71 this usage dates 
from the Tang period. Before that, lang alone also had the mean- 
ing of father ; 72 hence lang pa may be combination variant of 
lang and pa. In modern usage in that place, however, lang pa 
is always used in the referential and never in the vocative. 

Fu ch'in 3£IS 73 may be used as a literary vocative by the son 
in addressing his father, as in a letter. In this connection, ta jen 
JzA 74 and ch'i hsia WT 75 must be appended, making the term 
fu ch'in ta jen ch'i hsia, a stereotyped form of literary address. In 
addressing letters, ch'i hsia is used for either parent, whereas ta 
jen may be appended to any term for relatives of higher genera- 
tions. Ch'i hsia is primarily a literary parent term; its literal 
meaning is " like a child at your knees." 

The father, in referring to himself before his children, may use 
nai hung 75&, 76 nai wing Tim 77 and the more colloquial and 
modern terms a tieh, a pa, or lao tzu ^SiP. In certain localities 
lao tzu may also be used for father in general. 78 

The depreciatory term is chia fu, or chia yen. 79 Yen literally 
means " the stern and respected one." Chia chiin 80 is also fairly 
common. Chia kung 81 is a rather uncommon old term. 

"Hua yang chi, 1.13a: ffi, KW-fc. g &• ffl, IHfe WfPf^RBffi, * 

®mm- aufiJMH. ms#r£- . . . 

"SAu 7,1.9b: £Aii0g|W. Pri'AfcJftHW, 85.4a: H^SKBPSH 
;g:W> m+tia- gP*@5fc*M8Wg?. According to jfcO [cited by Ch'eng 
wei lu, 1. 26a] M^tlBIBO O g|MiL. 

78 Ch'in ^ means " relative," " parent." 

7i ^A literally means " big man," that is " senior." It is frequently used alone 
as a vocative, e.g., Shih Chi ^giftl^, 41. lib; ibid., ^M&%& 8.32a. 

"Hsiao Ching ffift£ f 5.4b: &§I£«TO#0# • • • Sfc »T. ffl» 

"Han Shu Bt«^«. '66.17a: *^*^, ^£|m«*T- B&3S#^ 

77 Jb^i is somewhat equivalent to 75-&- M« 75 is used in the sense of "your." 

78 Cheng tzu t'ung: s, v. fu, father. 

79 J Ching, 4.16a: ScA^lftS^, OO^lli^. ffwoo Ching MlB^> 5.1a: 

#**#**© ; 

80 Shih shuo hsin yii, lA. 3a: Heft" ■ Here the term cftia chiin is used both as a 
complimentary and depreciatory term. [From this point in Dr. Feng's article the 
editors have been forced to revert to their usual practice of providing only indispensable 
citations from the Chinese.] 

81 Chin shu 43. 6b. Chia kung is no longer used in this sense; it is now used for 
mother's father. 


The complimentary term is tsun ta jen #^cA. Tsun chun 82 
tsun kung 8Z tsun hou 84 jeng wing, jeng chun 85 in modern times 
are literary rather than vocative. 

K'ao % is a posthumous term. In ancient literature it was also 
used for the living father, being synonymous with ju 86 At present, 
however, it is primarily an epitaphic term. At different periods k'ao 
was used with various modifiers to express special circumstances. 
In the Chu Li of the Li Chi, huang k'ao 87 is used as a posthumous 
term for father, but in the Chi Fa of the same work it is used as a 
temple term for father's father's father. 88 Wang k'ao is a temple 
term in the Li Chi for father's father in connection with sacri- 
fices, 89 but during the Tang period it was occasionally used as a 
posthumous term for father. 90 Huang k'ao and wang k'ao have 
been prohibited for common use since the Yuan dynasty, being 
reserved only for the imperial family. 91 Thereafter the term hsien 
k'ao came into use as a universal epitaphic form. 92 But this term 
sharply contradicts the old usage, since in the Li Chi, hsien k'ao 
is used as a temple term for father's father's father's father. 93 

Fu chun tff^j is another popular epitaphic term. Originally, 
i. e., in the Han dynasty, only those who had been governors (t'ai 

83 Shih shuo hsin yii, 2 A. la; Chin shu, 75.5a. Sometimes chun chia tsun is used, 
e. g., Shih Shuo hsin yii, 2B. 34 a. 

83 Chin Shu, 82. lb; ibid., 92. 23b. 

84 Shih shuo hsin yii, 1A. 28b: H^. 

86 ^m an d S"S were used originally for those who received titles through sons 
who had risen to high official positions. Later they became common complimentary 

86 According to the Shuo Wen, k'ao means "old." Thus it could be applied to any 
old man. Its application to mean father is a later development and its use as a 
posthumous term is a still later specialization. Cf. Ch'in shu chi 1.5. 

67 Li Chi f 5.22a. 

88 Ibid., 46.8b-9a: J|#. 

"Ibid. 46.8-9: 3^. 

90 Ch'ang-li chi 24.10a. 

"This prohibition is best illustrated in the Yuan Hen chang TcAl^, 1908 edition, 

62 The term hsien k'ao j§J|5j was used for father much earlier than the Yuan period, 
e.g., Shu Ching (K'ang kao), 14,3a. It continued to be used down to the fourth 
and fifth centuries A. D. for parents both living and dead. Hsien means " great," 
" illustrious," etc. Down to the Sung period, hsien is predominantly used in the 
posthumous sense. Cf. also the Chin Shih Li, 5. 55b. 

83 H]II^ [Ch'in shu chi, 2. 3a] rather bemoans such contradiction. 


shou ^C^) could be called fu chiin by their sons, but since the 
Tang period the term has been used indiscriminately. 94 The 
ordinary posthumous terms are hsien fu $t%, wang fu t"3t , hsien 
ta fu 3fe^c-^, hsien chiin $cM ' , hsien tzu ^fc-p, hsien chiin tzu $fc 
It-?/ 5 hsien kung ^fc^, 0G etc. The complimentary posthumous 
terms are tsun hsien chiin #^fe^* and tsun fu #jfr. 87 
The ancient temple term for father is ?ii ffl. 98 

34. Mu # mother 

M u, like fu, is primarily a standard term and is seldom used in 
the vocative. Other ancient alternative terms are yii " and wen, 100 
but these terms can be applied to any old woman. During the 
T'ang period, niang tzu 101 was used for mother but at the same 
time it was also used for any young woman. This usage seemed 
to be northern. Niang tzu is used in modern terminology some- 
times as a husband's term for wife, sometimes for any young 

The most peculiar variant of the term for mother is tzu tzu 
fctfct which was used by the royalty of the Northern Ch'i dy- 
nasty ; 102 tzu being a term for older sister. According to the Shuo 
Wen, the people of Shu call mother chieh and the people of Huai 
nan call mother she. 10Z The older form of chieh £1, according to 
the Yii P'ien, is written Stfe. In the t^Uli'l of the Huai nan tzu, 
she is used for mother and Kao Yu comments that it is a Chiang 
Huai practice. 104 The Shuo Wen also states that shih $& was used 

94 Cf. Heng yen lu, 3. 3b-5b; Ctieng wei lu, 1.21a-b. 

95 Hsien chiin and hsien tzu originally were terms used during the feudal period by 
the nobility in referring to their deceased fathers. They became common terms at 
about the end of the period. 

90 Hou Han Shu 93. 15a. 97 Ch f ang-li chi 21. 6b. 

D8 Tso Chuan 32. 4b. Ni means " near," '* closer," i. e., the father is nearer than 
the father's father, etc. It is the same as §/g. Cf. also Shu Ching 10. 11a. 
90 $g, Shuo wen; Hsin Shu 3. lb. 

100 4£> Shuo wen; Kuang ya; Han fei tzu 10b. Also read ao, 

101 £Jl~F"> SAw /, 1.9b-10a; for its uses during the various periods, cf. Kai yii ts'ung 
kao, 38. la-3a. 

102 Pei Ch ( i shu 9. 4a. 

103 Hsu Shen seems to consider |tfc a variant of £§.. 

104 Huai nan tzu 16. 12a-b. Chiang Huai is the area between the Yangtze and 
Huai Rivers. 


for mother in Chiang Huai. Kuo Fo (267-324 A. D.) said that 
the people of Chiang Tung 105 called mother shih f^» also pro- 
nounced chih, or ch'ih. She, shih, chih and ctiih all seem to have 
been derived from the same root and probably represent variants 
of chieh. 106 Apparently, from the third century B. C. to the 
fourth century A.D., chieh, with its variant forms, was a very 
prevalent vocative for mother throughout the Yangtze valley. 
Even down to the thirteenth century A. D. mother was some- 
times called chieh chieh. 107 On the other hand, from Han to Tang 
times chia chia was frequently used for mother and might be 
another variant transcription of chieh chieh. 108 

Perhaps due to the close similarity of these two sounds — tzu 
and chieh — and to the vagaries of transcription, for a time tzu 
was used for chieh. Tzu, being the older and more literary term, 
triumphed over chieh, and the latter lost its original meaning and 
acquired the meaning of older sister, like tzu. This seems the only 
reasonable explanation, and, if true, means that we have here an 
exact parallel with hsiung and ko, as discussed above. Apparently, 
no marital relations are involved. 

The universal modern vocative is ma $M or ma ma. 109 Niang 
f&, 110 or niang niang, are also very commonly used in many 
localities. The Kuang Yun says that the people of Ch'u called 
mother ni 111 The Chi Yun states that the people of Ch'i called 
mother mi, 112 and the people of Wu called mother mi. 113 Ni and 

105 Chiang Tung is a vague geographical term, approximately the lower Yangtze delta. 

106 According to Cheng Chen (Ch'in shu chi, 1.6b) the ancient pronunciation of 
chieh $|L andflft s ^ was about the same and both were in the rhymes ^, ]^|» 
^| which are very close. Hence he considers she to be a dialectical variant of chieh. 

™Ssu ch'ao wen chien In HlSHUiJfc £j|, by Yeh Shao-weng MUm> 
ca, 1220 A. D. [£n2BPfU£j||# edition] 16a. 

108 Pei Ch'i shu 12.8b; %%. 

109 But cf. Hsi shang fu fan ;£ JJgfjfc, by Yil Yen #$ [%MlkM% edition] 

xo As explained before, niang can be used in various designations sometimes over- 
riding generations. Used as a vocative for mother, it was first noticed during the 
fourth and fifth centuries A. D. Nan Shih 44. 5a-b; Pei Shih 64. 13b-14a. 

111 i|f; cf. CKin shu chi, 1.6. Ch'u ^ is the ancient term for the middle Yangtze 
valley and at present approximately the modern Hupeh and Hunan provinces. 

112 |H- Cf . also the Yii p'ien. Ch'i is the old name for the modern province of 

113 il§- Wu is the ancient name for roughly the southern part of Kiangsu province. 


mi may be early dialectical variants of ma. The Miao, Yao and 
Tung tribes of Southwest China still call mother mi or ma. 11 * 
Whether the aboriginal terms influenced the Chinese, or vice 
versa, or whether they both have been derived from a common 
earlier form, we at present have no way to determine. Certainly 
ma is only a slightly differing version of mu. llb 

Mu ch'in "©IH is sometimes used as a vocative, but more com- 
monly in addressing one's mother in a letter; in the latter case, 
ta jen and ch'i hsia must be suffixed. 

The depreciatory term is chia mu 116 or chia tzu 111 Tzu lite- 
rally means " the affectionate one." Tsun lao 11B was used around 
the fifth century A. D., and chia fu jen was allowable during the 
Later Han times. 119 

Ling mu ^#, 120 ling tzu faWk, ling fang fal^t, and tsun fang 
HH*, 121 are the most common complimentary terms. Tsun shang 122 
and tsun fu jen 123 were used from the fifth to the eighth centuries 
A. D. At present tsun fu jen is used as a complimentary term for 
another's wife. Tai fu jen 124 may be used for another's mother 
when the father is dead. An jen i£cA and kungjen #A, originally 
terms for a titled woman, may be used loosely, if incorrectly, as 
complimentary terms. 

The posthumous term is pi 125 as defined by the Shuo Wen. Yet 
this view is sometimes disputed, since in classical literature the 
term was often used indiscriminately for both living and dead 

114 Miao fang pei Ian, op. cit., 8. 6a, and 9. 10a. Ling piao chi man, op. cit., 137. 

115 Ch'm shu chi, 1 . 7a. 

116 Yen shih chia hsiin, 2. 4b-5a. It seems that the prefixing of chia (house) to 
the terms of lineal ascendants to form depreciatory terms was not prevalent during 
Yen's time. 

117 This is the opposite of the term chia yen, " the stern one," for father. The 
mother is supposed to be affectionate and the father, stern. 

lls Sung Shu 91.15a: H^g. 

110 Hon Han Shu 78.9a: ^/^A- 

120 Ts'ai chung-lang chi 6. 7b. 

121 Lu shih-lung wen chi 10. 10a. Tang ^ is derived from pei t'ang ^bH^> a non " 
vocative, non-referential literary term for mother. Cf. Tung su pien, 18. 5a-5b. 

122 Sung shu 91.15a: ^_b- 

123 Ch'ang-li chi 29.3b: #^cA- 

124 This was originally used for a titled woman, e. g., Han Shu 4. 12a-b: >fc^A- 
126 ilt- The Erh Ya uses mu and pi synonymously. 


mother. The modern usage follows the interpretation of the Li 
Chi, that mu is used when the mother is living, and pi when she 
is dead. 126 Huang pi 127 was an old sacrificial term but has been 
forbidden since the Yuan period. Hsien pi 128 is exclusively an 
epitaphic term. 

35. Poiui&X fob 

The old term in the Erh Ya and / Li is shih fu. 129 In the Li Chi, 
po fu 13 ° is sometimes used in place of shih ju. Po itself means 
oldest, e. g., an oldest brother may be called po hsiung, and an 
oldest sister, po tzu. Since the Wei and Chin periods, po alone 
has been used as a vocative for father's older brother. 131 From 
the Sung period down to modern times, po po has been the most 
prevalent vocative. The posthumous term is wang po tTffi . Ts'ung 
hsiung ti men chung l32 is an old term used circa the fifth century 
A. D. but seldom heard today. 

36. Pomuf6# fobw 

Shih mu is the old term used in the Erh Ya, I Li 133 and Li Chi. 134, 
Po mu is also used in the Li Chi. 135 

37. Shufu&iUyb 

Circa the latter half of the first millennium B. C, chu fu, 13e 
tsung fu, 137 and yu ju 138 were used for father's brothers, both 

128 Li Chi 5. 22b. 127 Ibid. 5. 22a. 

128 Wang shih chung chi 38a: jgg This is, perhaps, the first use of hsien pi, but 
it is used for the living mother. Now hsien pi is used exclusively as a corresponding 
term to hsien k'ao. 

128 / Li 30. 8b. Shih means " generation." That is, the father's older brother is the 
one in the father's " generation " to succeed to the grandfather. 

130 Li Chi 18. 10a: f £ . 

131 See p. 197, note 89. 

132 Yen shih chia hsiin 2.6a: t3b^ IfrF 1 ! + • It literally means "within the gate 
of father's brother's sons/' a circumlocution for expressing a mournful situation. 

133 / Li 30.9b-10a. 

134 Li Chi 18.15a. 

135 Ibid., 43. 2b. 

136 Shih Ching 9C. 2a: ^0 . Chu fu h a very vague term, literally "the fathers." 
37 1 Li 31.17a: $JL0 Jjij^i. Since fathers brother's sons can be called ts'ung fu 

k'un ti, father's brothers can be called ts'ung fu. Pei shih 22. 8. 

138 Li Chi 8. 4b-5a: ?^0. Since brother's sons can be called yu tzu, it is inferred 
that father's brother can be called yu fu, " like father." 


older and younger. They are still used today as alternative terms 
but are primarily literary forms. Ts'ung fu, a contraction of 
tsung tsu fu, was also used for father's father's brother's sons. 

Shu fu also was used in another sense in early times. The 
father's first younger brother was called chung fu, the second 
younger brother, shu fu, and the youngest, chi fu. 1 ™ This usage 
was never common. Chi fu was also used for father's younger 
brothers in general, not necessarily his youngest brother. 140 Ts'ung 
weng is a relatively late and uncommon term. 141 

Since the third century A. D. shu, a shu, or shu shu have been 
used as common vocatives. 

Among the royal families of the Northern Ch'i and T'ang dy- 
nasties, father's brothers were called a hsiung; this may be a 
family peculiarity. 

The depreciatory term is chia shu fu, or simply chia shu. Yen's 
Family Instructions states that it is incorrect to use chia po for 
father's older brother, since he is an elder of father and one dare 
not use chia. 142 This is somewhat over-rationalistic; chia po is 
the most common form today. The complimentary term for an- 
other's father's younger brother is hsien shu, or ling shu. 

The posthumous term is wang shu. Ts'ung hsiung ti men 
chung is an old term both for father's older and younger brothers. 

38. ShumuWfybw 

This is also the term used in the Erh Ya. Chi mu 143 was used 
during the Han period, but rarely. The vocative shen 144 dates 
from the Sung dynasty. This term is not found in the classical 
literature, and is thought to be a contraction of shih mu. In modern 
times shen shen, shen mu, or shen niang have been commonly 
used as vocatives. Another common usage is by prefixing her sib- 
name, or the numerical order of her husband, to ma. This also 
applies to 36. 

189 Shih Ming: ^0. 

140 Shih chi 7. lb; Ch'ang-li Chi 23. 9b; Han Yu refers to himself as chi ju. 

141 T'ang chih yen, 3. la: ^tH . 

142 Yen shih chia hsun 2. 5a. 
1A *Hou Han Shu 118.20a. 

144 Ming tao tsa chih, 13b: $| . 


39. Kumu *£#fsi 

" Father's sisters are ku," defines the Erh Ya. Ku is also used 
in the / Li. 145 

The vocative for father's unmarried sister is ku, or ku ku pre- 
fixed by her name or numerical order. When she is married, ku 
ma is the most prevalent form. 

Chia ku mu is sometimes used as a depreciatory term. Yen's 
Family Instructions considered this usage incorrect, for when a 
woman was married out she was no longer a member of the 
family, so that chia could not be applied. 146 

40. Ku fu *fr£ f si h 

Ku hsiX 147 and ku fu 148 are used as alternative terms, mostly 
from the third to the sixth centuries A. D. 

41. Tangpofu £ffi3C ffbs>f 

Erh Ya gives the term ts'ung tsu fu for father's father's 
brother's sons, both older and younger than father. By inference 
from other usages, ts'ung tsu shih fu may be used for the former. 
Since the Han period, ts'ung po, or ts ( ung po fu have been used. 1 ' 49 
During the fifth and sixth centuries A. D. the term t ( ung fang was 
introduced, since agnate relatives of the same paternal grand- 
father offer sacrifices in the same ancestral hall. Tung t'ang lite- 
rally means " the same hall." During the T'ang period the t'ung 
was dropped and only t'ang was used. 150 Later the term was 
extended to other collateral lines. 

42. Tang po mu ^I6# w of 41 

As can be inferred from ts'ung tsu fu and ts'ung tsu shih fu, 
the older term would be ts'ung tsu mu or ts'ung tsu shih mu. 

145 I Li 31.16b-17a. Liang Ch'ang-chii (Cheng wei lu, 8. 13a-14b) considers ku 
tzu mei a term for father's sister. As far as its use in the I Li is concerned, it should 
be interpreted as ku (father's sister) , and tzu and mei (ego's own sisters) . The 
other instances which Liang cites in support of the use of ku tzu for father's older 
sister, and ku mei for father's younger sister, are very questionable. 

146 Yen shih chia hsiin 2. 5a. 

147 Pei Ch'i Shu 18.3b: jg. 

148 San Kuo Chih 13. 3a; Nan Shih 57. 14a. 

146 Chin Shu 80. la. 18 ° Ch'ien Ta-hsin [Heng yen lu, 3. 9a-b]. 


Since the Han period ts'ung po mu or simply ts'ung mu have 
been used. 

43. Tang shu fu ^&3C f f b s < f 

Compare 41. Tsung tsu shu fu may, by inference, be the older 
term. Since the Han period tsung shu or tsung shu fu 151 have 
been used as alternative terms. 

In the vocatives of 41 and 43 the modifier t'ang is usually 
dropped, i.e., the vocatives of 35 and 37 may be applied here, 
respectively, modified by their names or numerical order. 

44. Tang shu mu ^&# w of 43 

Ts'ung tsu shu mu may be the older form. The later abbreviated 
form ts'ung shu mu may be used alternatively with t'ang shu mu. 
T'ang shen is a vocative extension of shen. 

45. Tang ku mu ^f* f f bd 

The term in the Erh Ya is ts'ung tsu ku, later abbreviated to 
tsung ku. 152 Ts'ung and t'ang are synonymous. The vocative for 
father's sister may be used here, modified by her name or numeri- 
cal order. 

46. Tang ku fu £*fc£ f f b d h 

The vocative for father's sister's husband may be used here, 
modified by his sibname. 

47. Tsai ts'ung po fu Ht£f63£ s of 25 or 27 > f 

Tsu fu is the term used in the Erh Ya and / Li. 153 Tsu po fu 
may be used for this relation, but it can be applied to any male 
sib relative of the father's generation older than father, from the 
fourth collateral line onward — so it is a rather loose term. 

48. Tsai ts'ung po mu HtStffi^ w of 47 

49. Tsai ts'ung shu fu W&U5C s of 25 or 27 < f 

Tsu fu is used in the Erh Ya. Tsu shu 154 and tsung shu 1B6 are 
later alternatives, but used rather loosely. In the vocatives of 47 

161 Sung Shu 52. 5a, 

168 Chin Shu 51. 2b. 154 Chin Shu 83. la. 

153 / Li 33. 6a: j&. 1B8 Yin hua lu, %. 2a: £^. 


and 49, tsai tsung is usually dropped, leaving only po and shu 
modified by their names or numerical order. 

50. Tsai ts'ung shu mu H$t$t"fi£ w of 49. 

The vocatives of 36 and 38 may be applied to 48 and 50, 
respectively, modified by their sibnames or their husbands' nu- 
merical order, or by both. 

51. Tsai ts'ung ku mu W&i&-& d of 25 or 27 

In the Erh Ya, tsu tsu ku MM$& . In modern times, tsu ku 
^M has been used alternatively, but in a rather loose fashion. 

52. Tsai ts'ung ku fu WM&X h of 51 

53. Kupiaopofu M*f63C ff si s > f 

Chung wax chang jen 156 was used during the fifth and sixth 
centuries A.D. During the Sung dynasty piao chang jen 157 and 
wax po fu 158 were frequently used. Before the T'ang period, chang 
jen was used as a polite term for any old man ; but since then it 
has been used as a synonym for yo fu, wife's father. 

54. Ku piao po mu $i^f6# w of 53 

Chang mu 159 was used during the fifth and sixth centuries A. D., 
but is now used exclusively for wife's mother. 

55. Ku piao shu fu $MW£ f f si s < f 

56. Ku piao shu mu OOU^ w of 55 

57. Ku piao ku mu OOtt# f f si d 

58. Ku piao ku fu 00&3cl f f si d h 

59. Chiu piao po fu M$t1&H f mb s<f 

60. Chiu piao po mu OOf6# w of 59 

169 Ff^Kzfc A.' c ^- ^ en s hih chia hsiin 2. 8b. Chung-wai is synonymous with 

167 Tai ping kuang chi ^ZpJ^fg [193^ Peiping, 1 $C&'s£ edj 148.4a. 

168 cf. Mmmmmm rwMnc edj 5a . 

159 Yen Shih Chia hsiin 2. 8b. The use of wang 3E m ^ and hsieh H^ mu is no 
longer intelligible. Perhaps their use is based on the most well-known sibnames of the 
time, Wang and Hsieh. 


61. Chiu piao shu fu J!^t&3t f m b s < f 

62. Chiu piao shu mu 00&# w of 61 

63. Chiu piao ku mu OOM# fmbd 

64. Chiu piao ku fu OOM3t h of 63 

65. Ipiaopofu ^^f63C f m si s> f 

66. I piao po mu OOf&# w of 65 

67. I piao shu fu OOUH f m si s < f 

68. I piao shu mu CO&# h of 67 

69. I piao ku mu 00&# f m si d 

70. Ipiaokufu 00&3£ wof69 

The terms listed under 53 are applicable to 55, 58, 59, 61, 64, 
65, 67 and 70 respectively, during the period mentioned above. 
In ordinary modern usage the modifiers ku, chiu and i are usually 
omitted, so that terminologically these relatives are not distin- 
guished from one another. Conceptually, the exact relationship 
is always assumed. 

V. Generation of the speaker 

71. Pen shen ^# ego, a male. 

A female would use the same terms, except for those provided 
in Table IV and certain terms in Table III. 

72. Ch'i # wife 

Ch'i tzu 160 #-p is commonly used, but it may also mean " wife 
and children." In certain cases fu 161 is used synonymously with 
ch% but it may be used to mean " woman " in general. In kin- 
ship usage fu is principally used for the wives of those who are 
of lower generations and age status. 

180 Jih Chih lu, 24. 5a. 
101 7 Ching, 1.33a: jf. 


Fei 162 is a very old term and so also is nei c/m, 163 but the latter 
is rather uncommon. Nei she 164 and ju jen 1G5 were commonly 
employed during the first half of the first millennium A. D. 
Shift, 166 chia, 167 and shih chia 168 are also very old terms, but are 
still commonly employed as literary forms. 

Chieh fa 169 and chung k'uei 170 are primarily literary terms. 
Chieh fa is applied only to the first marriage principal wife ; chung 
k'uei is also used as a literary complimentary term. 

The wife, when speaking to the husband, calls herself ctiieh 
^, " your concubine " ; or, chi chou cJiieh. 171 These abject terms 
are seldom, if ever, used except in literature. During the feudal 
period, a noble woman could on formal occasions, call herself fei 
tzu, 172 or hsiao tung, 17 * according to rank. These terms are now 
entirely obsolete. Chun fu 174 is a common form, but mostly 
used in poetry. 

The husband calls the wife hsien ch'i, 175 or niang tzu, 176 and, 
more anciently, hsi chun} 77 These highly complimentary terms 
are rarely used in daily life, and are rather mere literary forms. 
CKing is a reciprocal term, i. e., used alike both by husband and 

182 M Erh Ya; Shuo Wen; Shih Ming, Tso chuan 5. 19a; Chan kuo tse 3. 83a. Fei, 
as used before the Han period, is a term for " wife " in general. It is only during the 
Han period and since that jei is reserved for the wives of nobility and the Emperor's 
secondary wives. Cf. Ctiin shu chi, 2. 5b-6b. 

163 Tso Chuan 42. 9b: |*J ^fc- Nei chu literally means " inside lord." 

164 *r$ : WiMMtikWrt [F« t'ai hsin yung chi, 1.13a]: ft &. 

165 Chiang Wen-t'ung chih 1.1b: fjHy^. Ju jen was originally a term for titled 
women during the feudal period, e.g., Li Chi 5.11b. It became a common term for 
wife during the third and fourth centuries A. D. Since then, it has reverted to its old 

166 / Li 6.8b: if. Li Chi 28.20b. 

167 Shih Ching 7B. 5b: M^^iM^- Tso chuan 14. 10b. 
108 Shih Ching 4A. 6b: ^^. 

160 Wen Hsiian, 29. 14b: ^ j§. 

170 / Ching 4.17a: 'f'M- Chung k'uei literally means "the family larder" and is 
used metaphorically. 

11 % Shih Chi 8.4a: JE^^c; i.e., a female with broom and dust-basket ready to 

172 Tso Chuan 15.2a: ^J^. 

173 Lun Yii, 16.10a: /]>#; Li Chi 5.12a. 

17i Shih Ching 13B.6a: f^f. 17 ° Pei ChH Shu 39.5a: j&^- 

*" j|ijj£ literally " virtuous wife." 177 Han Shu 65. 5a: Jffl^J. 


wife, and is now a predominately literary form of address. Hus- 
band and wife may call each other by name, or they may use no 
term at all and just refer to each other as " you," " he," and 
" she." After issue, teknonymy is the most common practice. 

The depreciatory term is nei. 17S Nei tzu 1*1-?, nei jen F^A 
and pi nei ffc^I are its derivatives. Cho ching 179 and shih jen 180 
are more literary. The above terms are used mostly in refined 
society. Chia li 18x and hsiang li 182 are vulgar terms. Hun chia, 1B * 
hun she, 18 * lao p'o, 165 and chia chu p'o 186 are vulgar and dialectical 
forms used mostly in the referential, and are not necessarily 

The complimentary term is fu jen ^A, originally a term for 
the wife of a man of rank. More intimately, sao fu jen £I^;A 
may be used. Tai t'ai is more colloquial. 187 Ling shih 188 and ling 
ch ( i 189 are literary in nature. In ancient times nei tzu 19 ° was 
used as a complimentary term, but is now exclusively used as a 

78 It is also a general term used for wife and concubine together, e, g., Tso Chuan 
14. 18b. Cf. Heng yen lu, 3. 9b-10a and Ch'eng wei lu, 5. 12b-13b. 

179 JJfiflJ is derived from the MWfc g* of Meng Kuang ^fcJt, wife of Liang 
Hung ^^ of the Later Han dynasty* The variants are shan eking \1\M> ching fu 
^lj#|r and ching jen $IJA- 

™Shih Ching 2C. 6a: gTA- 

181 "In the home." Cf. Yao K'uan JKtfS.Miei A. D., ffiggf&fg. *^J"i5 
ed., 2. 19b. 

182 *tBM> " m the country." Nan Shih 64. 16a: f|#|#, V^MMSL, I ft?f 

r8z Nan Tang chin shih l&fHill^ of Cheng Wen-pao ^^ff , 953-1013 A. D. 
l*KM1£ffl& ed.] 3a-4b: j?^; also Liang ch'i i kao ® gSfjftfH. f| # of Yu 
Mou ±H 1127-1194 A.D., [|§ Ol^RStf lj ed.] fltRlfc, lib. 

184 t^^" is a variant of hun-chia. 

85 ^§^| M old woman." Cf. Ch'eng wei lu, 5. 10a. This is the most common 
modern term. _ 

X8Q Heng <yen lu, 3. 13b-14a: f^fty?. 

187 Ho Liang-Chun fi«rfi.f£. Ssu yu chai ts'ung shuo E^if|||£ [%£§&j&B 
ed.] 6.52b-53a: Jc>Sw - Cf. also Tung su pien, 18.8a. During the Ming period, 
t'ai t'ai was a term for the wives of officials of the *p7& rank and higher. 

180 Shih Ching 20B,8b: >fr# 

190 Yen tzu cKun ch'iu, 6.9a: (^J-p. Nei tzu was used during the feudal period 
as a term for the wife of ministers of the feudal lords, e. g., Li Chi 44. 16a, and Shih 
ming. Hence it is sometimes used as a complimentary term. 


depreciatory term. K'ang li m is a literary form used to refer to 
another's wife, and, more commonly, to refer to both husband and 
wife as hsien kang li. 

Pin 192 and ling jen 193 are posthumous terms for wife. Te pei 194 
is a posthumous complimentary term. It is also used as a literary 
form for the wife in the case of an aged couple. 

73. Hsiung 5L o b 

The Erh Ya used hsiung to explain k wn. 195 The two terms 
apparently were synonymous in ancient times. Kuo PVs (276- 
324 A. D.) commentary on the Erh Ya states that the people of 
Chiang Tung called older brother k'un. The Shuo Wen does not 
give the character H", but gives the character IT, and states that 
the Chou people used this term for older brother. T is probably 
its original form and ^ and jfc, its variants. H is a later bor- 
rowed form. 196 In the Shih Ching, only the Wang Feng (4A. 9a) 
uses the term k'un, all others using hsiung. This fact is regarded 
as evidence that k'un is a Chou term. 

In the / Li all male paternal cousins of the same sibname of 
ego and within the Ta Kung degree of mourning are given as 
k'un ti, and all male paternal cousins of the same sibname but 
beyond the Ta Kung degree of mourning, and all male cousins 
of different sibnames from ego are given as hsiung ti. 191 This 
illustrates the intentional differentiation and standardization of 
the degrees of relationship adopted by the ritual books in employ- 
ing terms in other than their original connotation. K'un is 
entirely obsolete at present. 198 

191 Tso Chuan 27. 2b: {/lH ■ ™ Li Chi 5. 22b: Jg . 

193 Chu tzu nien p'u ^^^jffl by Wang Mu-hung 3EJKJfc 1668-1764 A. D. 

W«ft«»*3*A.u* ft A. 

194 Ch'eng wei lu t 5. 14a: WM- 
1Q *Erh Ya: f. 

198 The Shuo wen lists the character j| but not with the meaning " older brother." 
The Yu p'ien gives |f and states that it is the same as j|. Huang K'an M:$l 
[488-545 A. D.j says, that k'un j| means " bright," M brilliant." Out of reverence the 
older brother is called k'un (Ifeig-^Jft, 6.2b-3a, iZ^fflJtM edition). This is 
rather a rationalization. 

197 Ch'eng wei lu, 4. 2b. 

198 For a most complete and classical study of the difference between these characters, 
ci. SS&5£fc»J| by Tsang Yung *&£, 1767-1811, n&£$M »»3°]. 1. 


The vocative is ko 199 ?f or ko ko a euphonic duplication. Ko, 
as given in the Shuo Wen, does not mean older brother at all; 
it means " to sing/' or " a song." Ko was first used from the 
sixth to the eighth centuries A. D., to mean father and then in 
the ninth and tenth centuries A. D. it became a vocative for 
older brother. 200 

According to the Fang Yen, the people of Chin and Yang 
called older brother po or p'o 201 The Shih Ming says that the 
people of Ch'ing and Hsu used the term huang ^. Until the 
fourteenth to the seventeenth centuries A. D. the peoples of the 
lower Yangtze delta still called older brother $& huang. 202 #£ and 
M are pronounced about the same. The ancient pronunciation 
of hsiung 5t may have been huang $t, as the two characters are 
often used interchangeably. 203 The Pai hu t'ung uses $L to explain 


74. Sao lobw 

The Erh Ya reads, " Woman calls older brother's wife sao" 
but it does not give the man's term. Whether or not sao was 
exclusively a woman's term we have no means of knowing. The 
Shuo Wen defines sao as " older brother's wife " but does not 
specify man or woman; it was most probably a term for both 
man and woman. 20 * 

75. Ti J&yb 

The younger brother, when speaking to the older brother or 
sister, calls himself pi ti 205 or hsiao ti 206 The older brother or 
sister calls the younger brother hsien ti 207 These were old usages ; 
at present simply ti is used. 

109 Kuang yiin. 

200 Kai yii ts'ung k'ao, 37.25b. 

201 ■ 

1 Fang Yen, 10. 4a: /J9» Its pronunciation is very uncertain. ^^ 

Yen pei tsa chih ^Ht^iife by Lu Yu 1^^, ca. 1330 A. D. [#j3 % edition] 

203 Shih Ching, 3A.7a; ibid., 18B. la; Han shu, 76.6a, with Yen Shih-ku's note. 

204 For example, Chan kuo ts'e 3. 6a. 

205 San kuo chih 29.26a (commentary): MR^- 

206 *tli#:/J^. 

201 Shih Chi 86.9a: Jtlfj. 


Chia ti 20S was used as a depreciative from the third to the 
eighth centuries A. D. It is now incorrect to use this term. The 
present term is she ti, and the complimentary term, ling ti or 
hsien ti. 

76. Tifu %Mybw 

Ti hsi s&M is an alternative term. Fu and hsi are synonymous 
in relationship terminology. The older brother usually avoids the 
younger brother's wife, and vice versa. Conversation can be only 
formal, and a proper distance must be maintained. 

77. Tzu W o si 

Tzu is now used chiefly in standard and literary contexts. It 
may be doubled as tzu tzu. The universal vocative is chieh or 
chieh chieh. Compare 34 for changes of chieh. Chieh may also 
be used for any young woman, e. g., hsiao chieh, which is equiva- 
lent to " Miss." 

Hsu 209 was an ancient term for older sister, used in the state 
of Ch'u ; shao 210 was used in Ch'i. Ming 211 was an old term for 
father's concubine's daughter older than father's principal wife's 
daughter. Later, in certain regions, it was used to mean older 
sister in general. 212 Nil hsiung 213 may be used as a literary alterna- 
tive for tzu. 

78. Tzu fu Wfk o si h 

The term in the Erh Ya is sheng — a reflection of cross-cousin 
marriage. Tzu chang and tzu hsu 214 are modern alternatives. 
Chieh fu is nore colloquial. Teknonymy is the most common 
practice as regards the vocative. If there is no child, brother 
terms are usually used. 

The Erh Ya says that sisters call each other's husband ssu, 215 
" private " ; this is considered evidence of the sororate. It has 

808 Ts'ao tzu ch'ien cki p®iS^, Mb; Tang shu 162.22b. 
208 Li Sao [Wen Hsiian, 32. 9b]: ^ ; also Shuo Wen. 
210 Kuang yiin: tB . 
811 Tso chuan 2.2a: ^ '. 

* 12 Fang yen, 12.1a. S1S Shuo wen: ;£J£. 

214 Hou Han Shu 49. 12b: ; Chin Shu 39. 8a-b. [No. 18 on p. 150.] 
216 M*- Shih Ching 3B. 9a. The Shih ming gives the traditional, but rationalistic, 


long been obsolete. In modern times, brother and sister use the 
same term. 

79. Mei & y si 

Ti 216 is said to be a woman's term for younger sister but this 
is by no means certain. During the feudal period ti had a special 
connotation in connection with the yin marriage. Nil ti 217 is a 
literary alternative term for mei. The Shuo Wen states that the 
people of Ch'u called younger sister wei 218 which may be a variant 
of mei. 

The depreciatory term chia tzu is used for older sister, and 
she mei for younger sister. This use of these terms is continued 
even after the sisters are married, although theoretically this 
should not be done. 

80. Mei fu *fc* y si h 

In the Erh Ya the term is sheng — a reflection of cross-cousin 
marriage. The Erh Ya also says that sisters call one another's 
husbands ssu — a supposed reflection of the sororate. Mei chang 
and mei hsii are modern alternatives. For other usages, compare 

81. T'ang hsiung £5£ f b s > e 

Ts'ung fu k'un ti is the term in the Erh Ya and I Li 219 for the 
first male paternal cousins. Later it was abbreviated to ts'ung 
hsiung 22 ° for the father's brother's sons older than speaker, and 
ts'ung ti 221 for those younger than the speaker. Rung k'un ti 
was used in the Shih Chi 222 During the fifth and sixth centuries 
A. D. t'ung tang was substituted for ts'ung, e. g., t'ung fang 
hsiung and t'ung t'ang ti. During the latter part of the T'ang 

218 Shih Ming: $& . This interpretation is followed by Cheng Chen in his Ch'ao 
ching ch'ao wen chi, 1. 17b-18a. This view is hardly justifiable, however, on the 
basis of the Erh Yas use of mei and ti. The former is used both by man and woman; 
the latter, in reference to the yin custom of the feudal period, refers to younger 
sisters who have married the same man. Thus ti can also be applied to husband's 
younger brother's wife, as in the Erh Ya. 

217 Shuo wen: £j£. 

218 Jj|. Cf. also Kung-yang chuan 4.7b. 

219 / Li SI. 17a. 221 San kuo chih 8. la. 

220 Liang Shu 31. la. 222 Shih Chi 49. 6a: fc . 


dynasty the t'ung was dropped, only t'ang hsiung and t'ang ti 
being used. T'ang and ts'ung can still be used alternatively. 

82. T'ang sao £ft w of 81 

83. T'ang ti ^&fbs<e 

84. T'ang ti fu &%& w of 83 

85. T'ang tzu ^i* f b d > e 

The older term is ts'ung fu tzu mei for father's brother's 
daughters, both older and younger than speaker. The develop- 
ment is exactly parallel with that of 81, t'ang hsiung. 

86. T'ang tzu fu 005fc h of 85 

87. T'ang mei £&fbd<e 

88. T'ang mei fu 00=*c h of 87 

89. Tsai ts'ung hsiung W&R f f b s s > e 

The older term is ts'ung tsu k'un ti, as used in the £rh Ya and 
the I Zi. 223 Tsai ts'ung was substituted later. Tsai means " once 
again " or " a second time." Ts'ung is synonymous with the later 
term t'ang, which indicates the second collateral line. Hence tsai 
ts'ung indicated the third collateral line. 

90. Tsai ts'ung sao WftM w of 89 

91. Tsai ts'ung ti Bt£& ffbss<e 

92. Tsai ts'ung ti fu #ffis&# w of 91 

93. Tsai ts'ung tzu f$f£j$ f f b s d > e 

94. Tsai ts'ung tzu fu W&WJZ h of 93 

95. Tsai ts'ung mei Hfl-fle f f b s d < e 

96. Tsai ts'ung mei fu Hf£&* h of 95 

97. Tsu hsiung MR f f f b s s s > e 

The alternative and more exact term is san ts'ung hsiung =■%£. 

'" I Li 33. lb: 


5L. San means " third," and tsung signifies second collateral. 
Hence san tsung means the fourth collateral line, since tsung be- 
gins the count from the second collateral line. This principle can 
be extended and the terms formed, e. g., ssu tsung, wu tsung, 
liu tsung, indicating the fifth, sixth, seventh collateral lines, 
respectively. Tsu M is a vague term applied to all sib relatives 
from the fourth collateral line and beyond, without further dis- 
crimination. Tsung hsiung 224 is a modern alternative. Tsung ^ 
is in a certain sense synonymous with tsu M. Tsung tseng tsu 
k'un ti 225 was used during the Han period ; it is a rather clumsy 
device involving the enumeration of ancestors. 

Ch'in t ( ung hsing 226 is the term given in the Erh Ya for male 
paternal cousins of the same sibname, of the fifth collateral line. 
It is obsolete now ; generally tsu hsiung ti is used, or, more exactly, 
ssu ts'ung hsiung ti. 

98. Tsu sao WM w of 97 

99. Tsuti$e&fffbsss<e 

100. Tsu ti fu M^m w of 99 
101: TsutzuKJ* fffbssd>e 

102. Tsu tzu fu M&$z h of 101 

103. Tsumei»fffbssd<e 

104. Tsu mei fu ^i*r*c h of 103 

The vocatives for brothers and sisters, brothers' wives and 
sisters' husbands can be correspondingly applied to 81-104, in- 
dividualized by their names, numerical order, or sibnames. De- 
preciatory and complimentary forms can be formed in the regular 

105. Ku piao hsiung M^^t f si s > e 

The old term in the Erh Ya is sheng — a reflection of cross- 

224 Tit /L is just as indefinite as tsu hsiung, and can be applied to any older sib- 
brother from the fourth collateral line and beyond. But tsung hsiung was used during 
the feudal period by younger brothers to refer to the primogenitary eldest brother. 
Li Chi 19.10b-lla. 

226 Hsin Shu 8. 6a-b. 220 H^]^. 


cousin marriage. During the Han period wai hsiung ti 227 and 
ts'ung nei hsiung ti 228 were used for both the older and the 
younger. The term piao also dates from this period. Sheng, wai 
hsiung ti and ts'ung nei hsiung ti are all obsolete now. 

In the vocative ku is always dropped, leaving only piao hsiung 
or piao ko. Piao hsiung is more literary and formal, piao ko is 
strictly vocative. In certain localities lao piao 3£^ is used. 

106. Ku piao sao ifi^t w of 105 
Vocative piao sao. 

107. Ku piao ti tt^lfc f si s < e 

108. Ku piao ti fu OO^^t. w of 107 

109. Ku piao tzu OOW f si d < e 
Vocative piao chieh. 

lid. Ku piao tzu fu GO*** h of 109 

111. Ku piao mei OOM f si d < e 

112. Ku piao mei fu OOi*ffc h of 111 

113. Tang ku piao hsiung ^tt^ftft f f si s s > e 

114. Tang ku piao sao OO^t$? w of 113 

115. Tangku piao ti OO^:^ ff sis s<e 

116. Tang ku piao ti fu OOOIfc#f w of 115 

117. Tang ku piao tzu OOOW f f si s d > e 

118. Tang ku piao tzu fu OOO*** h of 117 

119. Tang ku piao mei OCO& f f si s d < e 

120. Tang ku piao mei fu COO*** h of 119 

Terms 113-120 may also be applied to the children of father's 
father's brother's, daughter. This is by inference only; no docu- 
mentary usage has been noted. 

* 27 / Li 33. 9b. 2a8 Wen Hsuan t 25. la. 


VI. Generation of the son 
181. Tzu ip son 

Tzu, in ancient times, was used to mean child, either male or 
female. Thus it was often compounded with other elements to 
signify son, e.g., chang fu tzu. 229 Erh 2 ™ is synonymous with tzu; 
it is now used mostly as a diminutive, with no sex connotation, 
so that it has to be combined with other elements to express son, 
as in the modern term erh tzu 2S1 Hsi 232 is an old term for son, 
but also has the indefinite meaning of child; consequently, the 
forms hsi nan 2ZZ for son and hsi nu for daughters are used. Ssu 
B? 234 means " descendant," and is also used for son. During the 
feudal period ssu tzu Me referred to the eldest succeeding son, but 
in modern terminology is used for the adopted son. Hsing?** in 
ancient times, may be used for son but it may mean any descend- 
ant, being synonymous with sheng, 2 * 7 " to bear." Nu 288 is another 
old term. 

Ku and ni M9 were uncommon old, perhaps local, terms for son. 
Tsai |£ 24 ° and tsai ffl 241 are modern dialectical forms, apparently 
derivatives from tzu. H may also be pronounced chien 242 

Tzu can be combined with various modifiers to express the more 

**° Shift Chi e7.19a-b: ^?3t; literally, "male child." 

1,0 Kuang ya: fj, . 

** l Shih Chi 52.2a. 

*" >U> means " to reproduce," or " to bear." Hence it is used both for male and 
female children. See note 233, and p. 239, note 255. 

*" Ts'ao tzu-chien chi 8. 1. 

"* Shu Ching 4. 7a. 

m Li Chi 4. 5a. 

*" Shih Ching 1C. 7a: £fe; and Tso Chuan 42. 32a. 

187 Since the ancient pronunciation of hsing is about the same as ^fe, they are used 
interchangeably. Shih Ching 20D. 7b. 

" 8 Shih Ching 9B. 10b: $*. 

- Kuang Ya: ft *fc. Cf. £&«. Ht%5t&. 6B. 4. 

140 Fang Yen, 10.1b. 

* 41 Cheng tzu t'ung, s. v. 

141 Cheng Chen (Ch'in shu chi, 2. 12b) says that the Chi Yiin gives the pronuncia- 
tion /LriMsU. This is incorrect and probably a confusion with ^. But chien may 
be a Tang pronunciation, e.g., in Hua yang chi, 1.13a. However, at present the 
character is pronounced differently in different localities. In Fukien, it is pronounced 
"chan," in Chekiang and Kiangsu, "Ian," and in Kiangsi, Kwangtung, Hupeh and 
Hunan, "tsai." 


exact and complicated relationships of sonship resulting from 
ancestor worship, inheritance, concubinage, divorce and remar- 
riage, adoption, etc. 

The son, in speaking to the father, calls himself nan M , " a 
male issue." Erh tzu 5riiP is more vocative, and nan is principally 
a literary form of address. During the mourning period the son 
refers to himself as ku tzu, 245 ai tzu, 24 * ku ai tzu, 245 pu hsiao nan, 24 * 
or chi jen 247 

The father calls the son erh tzu both in the vocative and in 
writing. In speaking, usually only the name is used. In writing, 
the relationship term is used together with the name, as " erh tzu 
so and so." This rule applies to relatives of all descending 

The complimentary term is ling lang. 248 Other combinations 
are lang chun 249 or hsien lang 250 Hsien tzu, 251 ling tzu, 252 and 
ling ssu 253 are alternative terms, somewhat more literary. Kung 
tzu 254 was originally a term for the sons of the feudal nobility 
and later for the sons of men of high official positions. But now 
it has become a general complimentary term almost as prevalent 
as ling lang. Another very common, rather vulgar, term is shao 
yeh &M, which also originally referred to the son of a man of 
rank or of an official, e. g., as used by the servants in referring to 
the master's son. 

The depreciatory term is hsiao tzu, or hsiao erh. More vulgarly 

jjA , " orphaned son/' used when mourning for the father when the mother 
is living. 

244 $k> " grieving son/' used when mourning for the mother when the father is 

245 This is used in mourning for either parent when both of them are dead. The 
differentiation began during the T'ang dynasty. Cf. Kai yii tsung k'ao, 37. 8a-9b. 

1 ^fC^J^ " unfilial son." 
r ShihChing 7B. 3b: |$A* 
§[5 is originally a title of office. During the Han period, high officials could 
appoint their sons lang. 4 Thus lang became a complimentary term. Cf. Ch'eng wei bu, 

2,0 Yii t'ai hsin yung chi, 1.18b. 
350 pT>t$il> Ssn pu ts'ung kan ed., 10.17a. 
251 Wei Wu-ti chih 43a-b. m Mo CM, 2. 12b. 

862 Nan Shih 59. 6a. 2 " Shih Shing 1C. 6b; ibid. 13A. 5a. 

246 ' 

247 ( 


hsiao ch'uan 'hit, M a little dog." Chien hsi 26e and jo hsi 25a are 
obsolete literary terms. 

122. Tzufu^Sfsw 

Hsi ju MM is a more colloquial term. Hsi M was originally 
written M, which means " son " or " child." During the Sung 
period the female classifier was added, forming t& 257 Thus it 
became a distinct term for daughter-in-law. Hsi ju may be used 
for the wives of all the relatives of descending generations. 

Son's wife is usually addressed with this name by her parents- 
in-law. The father-in-law sees her only on formal occasions, and 
usually maintains a proper distance. When the daughter-in-law 
gets older and has children, the parents-in-law may even use the 
grandchildren's term in referring to her — extreme extension of 

123. Nii *k daughter 

Nii tzu tzu 258 and ju jen tzu 259 are used in the Li Chi and / Li 
for daughter, in distinction to chang ju tzu for male child. Ying 260 
is said to have been an ancient term for daughter, but this is by 
no means certain. 

Nii erh is more colloquial. It is used both by parents and 
daughter, and is a general term as well. When the daughter writes 
to her parents only nii is used in referring to herself. 

The complimentary terms are ling yuan, 261 ling ai 262 nii hung 
tzu !&2^ and, more colloquially, ch'ienchin 263 and hsiao chieh 26 * 
Yu nii 265 was used during the feudal period, but is no longer used 

2 ^Shih Chi 43. 32b: $&• 

266 Nan Shih 46. 10a: ggjg.. 268 / Li 31. 16b-17a. Li Chi £. 13b. 

887 Cf. Ch'eng wei lu, 8. 17b-18a. 269 / Li 32. 4a. 

2ao Yu pien: §|* It is usually a term for infant. 

61 -TriS yuan is a term for a beautiful girl and ling yuan is probably derived from 
Shih Ching t 3A. 5a. 

263 'frjSr, your beloved one. 

363 ^f*#*> " a thousand taels of gold," that is precious. Yin -p'ao sui pi ^ffffegS 
|f, by Tsao Mou-chien fftftfg iZilfcWM edition] 8a. 

38 * /MJL was used during the Sung period as a term for young maid servants, or 
prostitutes. Cf. Kai yu ts'ung k'ao, 38. 12. At present it is used as a complimentary 
term for the daughter of another and for any young woman. 

295 Li Chi, 49.3a: 3t£, 


in this sense. Hsiao niang tzu was commonly used during the 
Tang and Sung periods. 266 

The depreciatory terms are hsiao nu, more vulgarly, hsiao ya 
t'ou /f» T M, " little handmaid." Hsi nil 267 is an old obsolete term, 
which could be used as a literary form, Chia tzu 268 was admissible 
during the Han period, but was never used in this .sense in later 
times and nowadays is used as a depreciatory term for older sister. 

124. Niihsii^dh 

Mencius uses the term sheng. The Erh Ya uses the term hsii, 
which may also mean " husband " in general. 269 Hsu is combined 
with a variety of qualifiers to signify daughter's husband, e. g., 
tzu hsii, 270 lang hsii, and hsii sheng 271 Other alternative terms 
are nii fu 272 and pan tzu. 273 Chiao k'o 27 * tung ch'uang? 7 * t'an 
ch'uang 276 k'uai hsii, chia hsii 277 and mi ch'in 278 are mostly 
literary forms, used more or less in a complimentary way. 

Ch'ing 279 was originally a dialectical form (Shantung) for 
daughter's husband, and later was commonly used as a literary 
term. The forms tsu pien ^ffi and ping shih ^$1 are erroneous 
derivatives of ch'ing. 280 Ch'ing is also used to mean " husband " 
in general. Ch'ing and hsii were originally complimentary terms 
for a man of ability. 

289 Cf. Kai yii ts'ung k'ao, 38. la. 

167 Shih Chi 8.4a: &£. 

" a Yen shih chia hsiin, 2.5a: ^0. [No. 10, p. 149.] 

" 9 Originally, it was a complimentary term for an able scholar. 

270 Shih Chi 89. 10a. 

871 (Ch'eng wet lu, 8.21b). [No. 14, p. 149.1 

278 Chin Shu 34. 8a. 

373 ^ir, half son. Liu pin-k'o Wn chi, #*$|, i&tfcWtiJBS^i, 10. 7. 

aT4 JSI^ literally, " delicate w or " graceful guest." It is non-vocative and non- 

,7B JICtW^ " th e one w h° occupies the bed in the eastern chamber," is based on the 
anecdote of 3E^^» %$fcBi]|S. Chin Shu 80.1b. Cf. Shih ch'ang t'an, 1.3b. 

277 $fc0 and f£Q mean practically the same thing. Pei Shih 34.16b-17a. 

278 Chiu Tang Shu 159. 7a: ffiM- 
™Fang Yen 3.1a: fpf. 

280 Fang Yen, loc, cit., commentary. Kuang Ya, loc. cit. 


The complimentary terms are ling hsii and ling t'an. Ling t'an 
is derived from t'an cKuang, and is rather uncommon. 
Depreciatory term, hsiao hsii. 

125. Chih & b s 

The Erh Ya gives no term for brother's son, man speaking. It 
is conjectured that brother's sons (man speaking) could be called 
sons, tzu. The Li Chi uses the term yu tzu, " like son," 281 but 
whether or not it is an established term is quite uncertain. Dur- 
ing the Han dynasty the term tsung tzu 282 was commonly em- 
ployed, but more commonly the purely descriptive forms hsiung 
tzu and ti tzu were used. There is evidence that brother's sons 
were simply called tzu. 285 

Chih, as used in the £rh Ya, was a woman's term for brother's 
son. It is similarly used in the / Li (32. lb) . The use of chih as 
a man's term for brother's son dates from the Chin period (265- 
420) . This usage originated in north China and then became 
general. 284 The woman's term for brother's son was then prefixed 
with nei, thus forming nei chih, in contra-distinction to chih. 

Chih nan is mostly a self -reference term. Yuan 285 is a compli- 
mentary term, not commonly used. The common complimentary 
term is ling chih. 

126. Chih fu ffl® b s w 

Chih hsi fu is more colloquial. As remarked above, hsi ju 
applies to the wives of all relatives of descending generations. 

127. Chih mi &ic b d 

Chih as used in the £rh Ya and / Li is devoid of sex connota- 
tion, a feature characteristic of one of a pair of reciprocal terms. 

881 Li Cfc. 8.4a. 

181 Skih shuo hsin yu, lA.21b (commentary): $£■"?'• 

888 Han sku 71.4a-b; Hou Han Shu 90B. 18b: t'sai Yung refers to his father's 
younger brother and himself as ju tzu. It also must be understood that whenever 
one is referring to well-known relationships, or in the vocative, the more inclusive 
terms fu tzu are usually used, otherwise the more exact terms. 

884 Yen skih ckia hsun 2. 7a. 

" 8 Wt as a complimentary term was based on the uncle-nephew relationship of $£%%£ 
and ESc^lt Cf. Shih shuo ksin yu, 3 A. 38a. 


Ku and chih are conceptual reciprocals. When chih was trans- 
formed into a man's term, it ceased to be reciprocal and the sex 
indicator was suffixed, e. g., chih nu for brother's daughter. The 
inverted form nil chih may also be used. Yu nil and tsung nu 
are the terms corresponding to yu tzu and ts'ung tzu. Hsiung nu 
and ti nil are descriptive terms corresponding to hsiung tzu and 
ti tzu. 

128. Chih hsu Mb dh 

J hsing 288 is an uncommon ancient term, rarely understood 
today. Hsiung hsii, " older brother's hsu" and ti hsii, " younger 
brother's hsii/ 9 are descriptive alternatives. Chih nil hsu is more 

129. Wai sheng #*4£ si s 

The Erh Ya gives the term ch £ u and, in a later passage, sheng. 
CKu is probably an older term than sheng, since sheng, but not 
ctiu, is used in the I Li. 287 Wai sheng came into use during the 
Chin dynasty; it is also written ^f*^. 288 Chai hsiang 289 is a term 
used, probably rarely, circa the first half of the first millennium 

130. Wai sheng fu fa$£W si s w 

131. Wai sheng mi ft$%ic si d 

132. Wai sheng hsii 00*§ si d h 

133. T'ang chih £S f b s s 

134. Tang chih fu OOJf f b s s w 

135. T'ang chih nu OO^C f b s d 

288 Ta tai li chi, 6. 7a: Mf& . 28T * U 33. 9a. 

288 Shih shuo hsin yu, 3A. 3a (commentary) . 

289 ^6^0* "house site," is of interesting origin. Chin shu 41.1a: Wei Shu was an 
orphan reared in his maternal grandmother's home, the Ning family. When the 
Ning's built a house, a geomancer prophecied that this house site, Chai hsiang, would 
have a daughter's son who would be great. Wei Shu's maternal grandmother con- 
sidered this prophecy had been fulfilled, when Wei Shu, although young, was brilliant 
and precocious. Wei Shu then said, " I will fulfill the prophecy of this good house 
site, chai hsiang." Pei Ch'i Shu 29.2b; Shih ch'ang t'an, 1.3a. 


136. T'ang chih hsii OO* f b s d h 

137. T'ang wai sheng £&» f b d s 

138. T'ang wai sheng fu 00»» f b d s w 

139. T'ang wai sheng nu OOS8£ f b d d 

140. T'ang wai sheng hsii OO^if f b d d h 

141. Ku piao chih M^jl f si s s 

142. Ku piao chih fu OQQM w of 141 

143. Ku piao chih nu OOO^T f si s d 

144. Ku piao chih hsii OOO^ h of 143 

145. Ku piao wai sheng OO&SB f si d s 

146. Ku piao wai sheng fu OO&SMJ w of 145 

147. Ku piao wai sheng nu OOOO^T f si d d 

148. Ku piao wai sheng hsii 0000*g h of 147 

149. Tsai ts'ung chih W&M s of 89 or 91 

150. Tsai ts'ung chih fu OOMM w of 149 

151. Tsai ts'ung chih nu OOOic d of 89 or 91 

152. Tsai ts'ung chih hsii OOGJf h of 151 
158. Tsu chih %*M s of 97 or 99 

154. Tsu chih fu 00» w of 153 

155. Tsu chih nu OO& d of 97 or 99 

156. Tsu chih hsii OOJf h of 155 

VII. Generation of the son's son 

157. Sun& ss 

Tzu hsing 29 ° is an old obsolete term. During the Chin period 

"°Shih Ching 1C. 7a; / Li 44.8b; Shih Chi 49.2a. 


wan sheng tt£, * late born," was used for son, likewise, hsiao wan 
sheng, < f little late born," for son's son. 291 Wen sun is a literary 
form derived from the Book of History; 292 it originally referred 
to King Wen's son's son. Sun erh and sun tzu are more colloquial, 
erh and tzu being diminutives. 

Chia sun 293 was used as a depreciatory term during the Han 
period, but has never been used since and is now considered in- 
correct. The correct depreciatory term is hsiao sun. 

Sun may be combined with various modifiers to express the 
exact relationships, e. g., chang sun for the oldest son's son, shih 
sun or cKeng chung sun 294 for the eldest son's eldest son, who 
must carry the three years mourning obligations in his father's 
place in the event that the father has died before the grandfather. 

158. Sun fu 1&M s s w 

159. Sun mi ffiit s d 

Sun, as in the / Li, may be used to mean grandchild or any 
descendant from the second descending generation and down. 
In modern usage sun nil is employed in contra-distinction to 
sun. The inverted form, nu sun, is also permissible. 

160. Sunhsu Mi$ sdh 

161. Wai sun ftM ds 

162. Wai sun fu 00*fr d s w 

163. Wai sun mi OO'k d d 

164. Wai sun hsii OO^ d d h 

165. Chih sun MM b s s 

Ts'ung sun is a term found in the Kuo Fti. 295 Yu sun 296 was 
occasionally used during the T'ang period, and earlier, but is 
seldom used today, except as a literary form. 

191 Chin ahu 69. 7b; Fieh chi, 4. 2a. 

***Shu Ching 17.35a: %&. 

,ta Yen shih chia hsiln 2- 5a. ™ Kuo yu 3.7a. 

894 / Li 30. 12a: ^jjl {ft jfcMM- m Yiian * hih Chang-ch'ing chi 54. 4b. 


166. Chih sun fu 00# b s s w 

167. Chih sun mi 00£ b s d 

168. Chih sun hsii OO* b s d h 

169. Wai chih sun ft-BM b d s 

170. Wai chih sun fu 000*1 bdsw 

171. Wai chih sun nii 000£ b d d 

172. Wai chih sun hsii OOO^ b d d h 

In local variations, chih wai sun MRU and fang wai sun 
lkfl-%k may be used for 169-172. 

173. Wai sheng sun #>S}3fc si s s 

The term in the Erh Ya is li sun ffeU, literally "departing 
grandson." Whether or not there is any significance in this term 
one cannot say. Other ancient alternative terms are ts'ung sun 
sheng 2 " and mi sun. 23 * 

174. Wai sheng sun fu fl-S^if si s s w 

175. Wai sheng sun nu OOUic si s d 

176. Wai sheng sun hsii OO^^f si s d h 

177. Tang chih sun %MM f b s s s 

178. Tang chih sun fu 000*f w of 177 

179. Tang chih sun nu OOOiK f b s s d 

180. Tang chih sun hsii OOO^ h of 179 

181. Ku piao chih sun $3%$M f si s s s 

182. Ku piao chih sun fu OOOOfll w of 181 

183. Ku piao chih sun nu OOOO^T f si s s d 

184. Ku piao chih sun hsii OOOO^f h of 183 

185. Tsai ts'ung chih sun W&MM, s of 149 

m Tso Chuan 60. 20b. » s Ibid 60. 17b: ST 


186. Tsai ts'ung chih sun fu OOOOJt w of 185 

187. Tsai ts'ung chih sun nii OOOO^C d of 149 

188. Tsai ts'ung chih sun hsii OOCOf h of 187 

189. Tsu sun WM s of 153 

190. Tsu sun fu OOflfr w of 189 

191. Tsu sun nii OOic d of 153 

192. Tsu sun hsii OO^ h of 191 

Tsu chih sun W$£M may be used in substitution for tsu sun in 
terms 189-192, but the chih is not necessary. 

VIII. Generation of the son's son's son 

193. Tseng sun #3$ s s s 

According to old usages all descendants from the son's son's 
son and descending can be called tseng sun, or hsi sun. 298 * During 
the Han period erh sun was probably used synonymously with 
tseng sun, 299 

CKung sun £#> is the modern colloquial term. 

194. Tseng sun fu ft&M s s s w 

195. Tseng sun nii itMit s s d 

196. Tseng sun hsii OO 1 ^ s s d h 

197. Wai sun tseng sun ftMitM d s s, or s d s 

198. Wai sun tseng sun nii OOOO^C d s d, or s d d 

199. Tseng chih sun itMM b s s s 

200. Tseng chih sun nii OOO^C b s s d 

* 98 *Chm Tang shu 160. 19b. 

*" Han Shu 2.2b. The interpretations of the term erh sun 5|f, are quite 
divergent. Perhaps the interpretation of $Ui is the more prevalent usage in the 
Han period but it by no means precludes its use in the other connotations. Cf. Hsueh 
Lin S. 10-11. 


201. Wai sheng tseng sun MM&U si s s s 

202. Wai sheng tseng sun nii OOOO^T si s s d 

IX. Generation of the son's son's son's son 

203. Hsiian sun %M 300 s s s s 

204. Hsiian sun fu 00#t s s s s w 

205. Hsiian sun nii OOic s s s d 

206. Hsiian sun hsu 00*S s s s d h 

The following terms are found in the Erh Ya; although of no 
practical use, they are given here because of their theoretical 

207. Lai sun &M s s s s s 

208. K'un sun HO s s s s s s 

209. Jeng sun WO s s s s s s s 

210. Yiin sun HO ssssssss 


Relatives Through Mother 
I. Generation of the mother's father's father 

1. Wai tseng tsu fu &#i&3t mf f 

In the Erh Ya the term is wai tseng wang fu 003E3(I . 

2. Wai tseng tsu mu OOM# m f m 

Wai tseng wang mu OQ3E#, as used in the Erh Ya. Vocatives 
of the above two terms vary locally; they are largely based on 
the vocatives of 3 and 4, with generation modifiers. 

II. Generation of the mother's father 

3. Wai tsu fu #*ffi:£ mf 

Wai tsu may be used alone. The term in the £rh Ya is wai 

* 00 JihCkihlu,5.$2b. 


wang fu &5l£. Wai ta fu l and wai weng 2 are modern alterna- 
tives. The modern vocatives chia kung? also pronounced ka kung, 
and wai kung were used as early as the fifth century A. D. 

4. Wai tsu mu ft-I^M m m 

Wai wang mu is used in the Erh Ya. Wai po* is the most 
common modern vocative; likewise common is chia p'o, also 
pronounced ka po. Chia mu 5 was used during the fifth and 
sixth centuries A. D. Since chia at that time meant u mother," 
chia mu meant mother's mother. Chia mu is now used as a 
depreciatory term for mother. Liao liao* is a dialectical form 
used in certain parts of North China. 

5. Wai po tsu fu ^ffiffi^ m f o b 

6. Wai po tsu mu 000# m f o b w 

7. Wai shu tsu fu ftMWill m f y b 

8. Wai shu tsu mu OOQ# m f y b w 

Ku wai tsu mu may be used for the mother's father's sister, 
chiu wai tsu fu for the mother's mother's brother, and i wai tsu 
mu for the mother's mother's sister. These relationships are not 
maintained socially, but the terms show how they can be handled. 
Practically, there are many ways of solving this terminological 
problem, e. g., if the necessity for addressing these relatives should 
arise, one may adopt the terms of the mother's brother's son, who 
is the nearest relative of the same generation of ego on the mother's 

1 Chang Yu-shih wen chi W&$t3CM> collected works of Chang Lei »jj|^ 
(1052-1112 A.D,), Ssu pu U'ung Van edition, 17.9b. 

a Yuan shik chang-ch'ing chi 9. 5b. 

8 Yen shik chia hsiin 2. 8a. 

* Jung chat sui pi, ssu pi 2. lia. 

B See note 3. 

'K'ang-hsi tzu tien j$. Liao is also read lao, and is synonymous with ao. 
Yen Chih-t'ui [Yen shih chia hsiin, 2.7b] says that during his time the uncultured 
people called mother's parents by the same terms as father's parents when the latter 
were all dead. 


IIL Generation of the mother 

9. Chiu fuJi^mb 

Or simply chiu, as used in the Erh Ya. Since in modern usage, 
chiu also means wife's brother, the generation and sex indicator 
fu is necessary. Po chiu f6 Jl may be used for the mother's older 
brother, and shu chiu &Jt for the younger brother. These terms 
are now mainly literary. The mother's brothers and their family 
may be vaguely referred to as wax shih or wai chia. 1 

Vocatives are chiu chiu 9 a chiu, or chiu tieh. Chia chiu 8 was 
used as a depreciatory term during the fourth and fifth centuries 
A. D., but is not used today and is considered incorrect. Ling 
chiu and tsun chiu are complimentary terms. 

10. Chiu mu Ji# m b w 

Chin 9 is an old vocative used during the Sung period, and is 
now rather uncommon. The modern vocative is chiu ma. 

11. I mu $£# msi 

In the Erh Ya and the / Li the term is tsung mu. 10 I originally 
meant " wife's sister." The first use of i for mother's sister is 
found in the Tso Chuan, the 23rd year (B. C. 550) of Duke 
Hsiang. 11 This extension is attributed to the psychological simi- 
larity of these two relatives and to reverse teknonymy. Since the 
Han period i has entirely displaced the older term tsung mu. I is 
also used for concubines — a usage attributed to the sororate. The 
inverted form mu i may also be used. 

Vocatives are a i or i ma. 

W. I fu W£ m si h 

^ I fu, 12 i chang jen™ or i chang 9 were used during the first millen- 
nium A. D., but are uncommon today. 

'Literally "outer family." Chin Shu 41.1a. 

8 Shik shuo ksin yu, SB. 32a. 

9 Shu I, 1.9b: ^. Ming t'ao tsa chih, 13b. 

10 / Li 33. 2a. 

11 Tso Chuan 35. 18a. 12 Yen shih chia hsiin 3. 23a. 

" Pei Shih 47. 7b-8a. / chang jen is now used to mean wife's mother's sister's 


13. Tang chiu fu HJtX m f b s 

Tsung chiu is used in the Erh Ya and can still be used today, 
but more as a literary form. 

14. Tang chiu mu H*Ji# w of 13 

15. Tang i mu ^^# m f b d 

16. Tang i fu £«£ h of 15 

IV. Generation of the speaker 

17. Chiu piao hsiung Ji^Jt m b s > e 

Sheng is the term in the Erh Ya — a reflection of cross-cousin 
marriage. Nei hsiung 14 was a substitute used during the Han 
period; later it became confused with wai hsiung, 15 a term for 
father's sister's son. Today these terms are not used with the 
same meanings. Nei hsiung is now used for the wife's older 
brother, and wai hsiung for half brothers, by the same mother, 
older than ego. Chiu tzu 16 and chiu ti 17 are purely descriptive 
terms which were used circa 500 A. D.; at present they both 
mean " wife's brothers." 

Piao was first introduced during the latter part of the Han 
period, after cross-cousin marriage had long ceased to be prefer- 
ential. Piao, or chung piao, ls was first used for mother's brother's 
and father's sister's children, and later was extended to include 
mother's sister's children. 

In the vocative chiu is always dropped, leaving simply piao 
hsiung or piao ko, and, in certain localities, lao piao. 

18. Chiu piao sao Jl^tt w of 17 

19. Chiu piao ti OQ& m b s < e 

20. Chiu piao ti fu 00%*M w of 19 

21. Chiu piao tzii OOtft m b d > e 

22. Chiu piao tzu fu OOJ*:5*c h of 21 

14 1 Li S3. 10a. 18 Chin Shu S4. 8b. 

15 Sung shu 93. 4b-5a. 1T Ch'ang-li Chi 32. 7b. 

18 San kuo chih 11.20. Chung jriao is equivalent to nei wai. 


23. Chiu piao mei OO& m b d < e 

24. Chiu piao mei fu OO&^c h of 23 

25. I piao hsiung M^5L m si s > e 

Ts'ung mu leun ti is used in the Erh Ya and I Li (33. 9a) . J 
hsiung ti was used during the last few centuries of the first millen- 
nium B. C. and approximately the first half of the first millennium 
A. D. 19 It can still today be used as an alternative term. 20 Wai 
hsiung ti was sometimes used during the Tang period, 21 having 
resulted from a confusion with the term for mother's brother's 
sons and father's sister's sons, which finally led to the extension 
of the term piao and the partial merging of the three relationships. 

26. I piao sao ^fe#g w of 25 

27. Ipiaoti OOlfe msis<e 

28. I piao ti fu 00%m w of 27 

29. I piao tzu OOJ# m si d > e 

30. I piao tzu fu CO*^ h of 29 

31. I piao mei COM m si d < e 

32. I piao mei fu 00&f*c h of 31 

33. Tang chiu piao hsiung ^Ji^tft m f b s s > e 

34. Tang chiu piao sao OOOSS w of 33 

35. T'ang chiu piao ti OOOIfc m f b s s < e 

36. Tang chiu piao ti fu OOO&Jf w of 35 

37. Tang chiu piao tzu ^M*ft* m f b s d > e 

38. Tang chiu piao tzu fu OOOO* h of 37 

19 See discussion, pp. 177-178. Nan shih 57,17a. 

20 Cf. Liang Chang-chu, Ch'eng wei lu, 3.20b. Wang Shih-han £Eft]i$f:, 1707-? 
A. D., considers i hsiung ti a northern peculiarity: Tan shu lu SfcHH, (HSf^JHf 
ed.) 45a. 

I (cited by Ch'eng wei lu t 3. 20a) . 


89. Tang chiu piao mei HJt^M m f b s d < e 

40. Tang chiu piao mei fu OOOO* h of 39 

Terms 33-40 may also be applied to the children of father's 
mother's brother's sons, i.e., the children of 59-62 in table I, 
hence these terms are not given there. This extension is inferred 
only from popular usage, there being insufficient documentary 
evidence for collation. At any rate, when one considers non-sib 
relatives on the third collateral line in as far as the third gen- 
eration descending, the terminology becomes vague, and, indeed, 
an accurate system is not necessary here, since in almost all 
cases these relationships are not maintained socially. 

41. Tang i piao hsiung lEMMR m f b d s > e 

42. Tang i piao sao OO^Jft w of 41 
48. Tang i piao ti 00^& m f b d s < e 

44. Tang i piao ti fu OO^Ife*! w of 43 

45. Tang i piao tzu ^^^t* m f b d d > e 

46. Tang i piao tzu fu OO^J*^ h of 45 

47. Tang i piao mei CO^& m f b d d < e 

48. Tang i piao mei fu OO^fc^c h of 47 

V. Generation of the son 

49. Chiu piao chih M&M m b s s 

50. Chiu piao chih fu OOOJf mbssw 

51. Chiu piao chih nil OOO^C m b s d 

52. Chiu piao chih hsii OOCO*# m b s d h 

53. Chiu piao wai-sheng S^fl*SJ m b d s 

54. Chiu piao wai-sheng fu OOOOif mbdsw 

55. Chiu piao wai-sheng-nii OOOOic mbdd 


56. Chiu piao wai-sheng hsu OOOO^ mbddh 

57. I piao chih $&MM m si s s 

58. I piao chih fu OOOJt m si s s w 

59. I piao chih nii OOO^T m si s d 

60. I piao chih hsu OOQ*f m si s d h 

61. I piao wai sheng OO^-Sr m si d s 

62. I piao wai-sheng fu OOOOJf m si d s w 

63. I piao wai-sheng nii M%:fr$Bic m si d d 

64. I piao wai-sheng hsu ®@®@*f m si d d h 

65. Tang chiu piao chih 1kM^$£ s of 33 or 35 

66. Tang chiu piao chih fu ©©©©it w of 65 

67. T'ang chiu piao chih nii ©©©©;& d of 33 or 35 

68. Tang chiu piao chih hsu ©@®©ig h of 67 

69. Tang i piao chih Ik^^SL s of 41 or 43 

70. Tang i piao chih fu ©@©©*ff w of 69 

71. T'ang i piao chih nii ©©©©:& d of 41 or 43 

72. Tang i piao chih hsu ©®®@ig h of 71 

VI. Generation of the son's son 

73. Chiu piao chih sun M^MM m b s s s 

74. Chiu piao chih sun fu ©®@©Ji m b s s s w 

75. Chiu piao chih sun nii ©@©©ic mbssd 

76. Chiu piao chih sun hsu ®®©@*ir m b s s d h 

77. I piao chih sun %$%.&M m si s s s 

78. I piao chih sun fu ®©©©Jf m si s s s w 


79. I piao chih sun nii ®@®©^C m si s s d 

80. I piao chih sun hsii ©®®®*S m si s s d h 

Since most of these terms are extensions from table I, their 
historical development, vocative and complimentary usages can 
be inferred from there. 

Affinal Relatives — Table III 
Relatives Through Wife 
I. Generation of the wife's father 

1. Yofu «£wf 

The term in the Erh Yd is wai chiu, and in the Li Chi (52. 
27b) , chiu is used alone for this relationship. As chiu also meant 
mother's brother during this period, the terminology reflects 
cross-cousin marriage. During the Later Han period fu hung x 
and fu wing 2 were prevailingly used, as purely descriptive terms. 
Whether or not chang jen 2t A was used for wife's father during the 
Han period is by no means clear ; 3 it became the prevailing term 
during the T'ang dynasty. 4 Yo fu and chang jen are the universal 
modern terms, and the use of the one or the other depends upon 
local custom. Yo fu is more formal and literary, chang jen is 
more colloquial. Sometimes the combined and abbreviated form 
yo chang is used. Another very common but non- vocative and 
non-referential term is t'ai shan, name of the eastern sacred moun- 
tain of the old Chinese Empire. There are many interpretations 
of the origin of the terms yo fu, chang jen and t'ai shan, of interest 
to those who are concerned with the origin of individual terms .^ 

One interpretation of the origin of the term yo is that found in 
the Chiao ssu chih of the Han Shu (25 A. 13a) , viz., large mountains 
are called yo shan, and small mountains, yo hsii. Since mountains 
can be called both yo and hsii, and since hsii also means daughter's 

1 Hou Han Shu 71. Ha. 2 San Kuo Chih 1. Mb-ftSa. 

3 Han Shu 94A. 25b-26a. In the Neng kai chia man Ju, 2.28, this erroneously con- 
sidered to be the origin of chang jen for wife's father. The term chang jen used here 
merely means any older man; cf. Yen Shih-ku's commentary on this passage. 

4 Chiu Tang Shu 147. lb. 


husband, the meaning of yo was transferred and yo became a term 
for wife's father. 5 Another interpretation goes thus: Yo Kuang 
of the Chin dynasty was the father of the wife of Wei Chieh; 
since these two men were the best-known personages of their time' 
and since their relationship as father-in-law and son-in-law was 
much admired by the people, it is possible that yo chang ®3t 
is a corruption of Yo chang $k£. 6 

The story of the origin of the term t'ai shan is as follows: In 
the year 725 A. D. the Emperor Hsuan-tsung offered sacrifices to 
T'ai Shan, " Mount T'ai." According to precedence, all those 
officials^ who participated in it, with the exception of the San 
Kung — <&, were promoted one rank. Chang Yiieh, the premier, 
was the marshal of ceremonies. This son-in-law, Cheng I, was 
promoted from the ninth rank to the fifth rank and was accorded 
the privilege of wearing purple robes. In the banquet of cele- 
bration the Emperor was surprised by his quick advancement. 
The professional court jester, Huang Fa-cho, remarked: " This 
is the influence of T'ai Shan! " This is popularly considered the 
explanation of the origin of the term. 7 But t'ai shan must al- 
ready have had the meaning of father-in-law, since this joke is a 
pun, t'ai shan being interpreted both as wife's father and as 
Mount T'ai; i.e., Cheng I's unprecedented promotion was due 
to his participation in the sacrifices to T'ai Shan, or, in a satirical 
sense, to the influence of his father-in-law, Chang Yiieh. 

Still another version of the origin of the terms t'ai shan and yo 
relates that Mount T'ai, also called Eastern Yo, has a peak named 
Chang Jen. Since chang jen means wife's father, and Chang Jen 
is one of the peaks of T'ai Shan, t'ai shan has become a term for 
wife's father— a kind of punning and semantic transference. 
Furthermore, T'ai Shan is also called Yo, whence the term yo 
is derived. 8 

These are interesting speculations, any one of which is just as 
reasonable as any other. One point seems to be certain, viz., that 

J h Jih mn Chai V chi H *!3§F*f2 ^ mW [1277-1S57 A. D.] [ mm&m edition] 

' Chin Shu 36. lSb; Kai yu U'ung k'ao 37. 20a. 

T Cf . Shih ch'ang fan 1 . 2b. • Kai yu U'ung k'ao 37. 20a. 


no sociological factors or marital implications are involved. First, 
from the linguistic point of view, both yo and t'ai shan have never 
been used in any sense other than " venerable high mountain " 
and " Mount T'ai." Second, as relationship terms, both of them 
are late introductions, not earlier than the Tang period. If there 
were any sociological implications, they should be easily detectable. 

The application of chang jen to wife's father, as remarked above, 
first became prevalent during the T'ang period. Before and dur- 
ing the Han periods it could be applied to any old man to whom 
one might wish to pay respect. From the fourth to the sixth 
centuries A. D. chang jen was used for mother's brother, mother's 
sister's husband and father's sister's husband, e.g., chung wai 
chang jen. Hence, its use for wife's father may be an alternative 
extension from the term chiu, which was used during this period 
for mother's brother and sometimes for wife's father. If this idea 
is correct, the use of chang jen may be an indirect survival from 
cross-cousin marriage. 

Wai fu* ping sou, 10 and ping wing, 11 were alternative terms 
used during the Sung period. Fu t'o 12 is an old dialectal term 
used in southwest China during the Han period. 

Chia yo is used as a depreciatory term, but theoretically it may 
be incorrect. 

2. Yo mu &$k w m 

Wai ku or ku alone are used in the Erh Ya and Li Chi — a re- 
flection of cross-cousin marriage. Chang mu and t'ai shui 1S are 
terms corresponding to chang jen and t'ai shan. Before the T'ang 

•Cf. Ch'ien chu In $|K3$ l^huo Fu tftftf, 32] lb. 

10 Tung po ch'iian chi jft^^^, Ssu pu pei yao ed., 13.7: 7jC]§- 

11 Yu huan chi wen #^JjSffl, by Chang Shih-nan ^git^ (ca. 1200 A.D.), 
^H^JE^II^^ 6. 2b. On ping sou and ping wing see p. 255, note 6. 

12 Fang Yen, 6.7a: 0#^. 

» 'fril^gi (cited by the Ch'eng wei lu, 7.13a): JJItJC- The term t'ai shui is 
really interesting. The opposite of shan, mountain, is shut, water. Water is here used 
in the sense of " rivers " or " lakes." Since wife's father is called t'ai shan, so t'ai 
shui is used for wife's mother. The Ho pi shih lei being a cyclopaedia compiled during 
the Sung period, the term must have been quite common during that time. At 
present it is not a very good term and employed mostly on non-vocative and non- 
referential occasions. 


period, chang mu could be applied to father's and mother's mar- 
ried sisters, mother's brother's wife, or the wife of any person 
whom one addressed as chang jen. Mu t'o 14 was a dialectal form 
corresponding to ju t'o. 

Yo ju and yo mu may be used vocatively, but generally the 
husband adopts the wife's terms, addressing her parents with 
parent terms. Post-issue, teknonymy is the most common practice. 

In referring to wife's father's parents, circumlocution by enu- 
meration of relations is common. In certain localities lao chang 
jen and lao chang mu are used in the referential. Inferentially and 
logically, yo tsu ju and yo tsu mu would be correct, but they are 
not used. In the vocative, one usually adopts the wife's terms. 

3. Po yo fu {&&X w f o b 

4. Po yo mu f6lfr# w f o b w 

5. Shuyofu %L&X wf yb 

6. Shu yo mu &-&# w f y b w 

Alternatively and more commonly, po chang jen is used for 3, 
po chang mu for 4, shu chang jen for 5, and shu chang mu for 6. 
Lieh yo 15 is an uncommon complimentary term for 3 and 5. 

Wife's father's sister and her husband are called ku chang mu 
and ku chang jen, wife's mother's sister and her husband are 
called i chang mu and i chang jen, and wife's mother's brother 
and his wife, chiu chang jen and chiu chang mu, respectively. 

II. --Generation of the wife 

7. Chiu hsiung M£ wob 

Sheng, which is used in the Erh Ya, reflects cross-cousin mar- 
riage. The Erh Ya also gives the term hun hsiung ti, " brothers 
by marriage " a purely descriptive term. The Li Chi 16 gives the 
term ssu ch'in hsiung ti, also more or less descriptive, since ssu 
ch'in literally means " private relations." Fu hsiung ti 1T and nei 

14 Cf . Fang yen 6. 7a. 

15 Ho pi skik lei (cited by Ch'eng wei Lu 7. 13a), ?!j^. 

M Li Chi 27. lib. " See p. 195, note 84. 


hsiung ti 18 were in use from the Chin to the Tang dynasties, and 
can still be used as alternative terms. Fu and nei both mean 
" wife." Chiu was first applied during the tenth century A. D., 
through teknonymy. 19 

8. Chiu sao Jl#§ w o b w 

CKi sao 20 may be used, but is a purely descriptive term. 

9. Chiuti jlffe wyb 

10. Chiu ti fu M%W wyb w 

11. I tzu Wfc wo si 

12. I tzu fu bMWJi w o si h 

The term in the Erh Ya for wife's sisters' husbands is ya 21 or 
yin ya as used in the Shih Ching 22 Yu hsii 23 was used during 
the Han period and t'ung men a little later. 24 Liao hsii originated 
as a local term in eastern China. Lien mei and lien chin 25 were 
first used during the Sung period. Lien chin is the most commonly 
used term at present; it probably originated as a local form in 
North China. J fu, 26 as a term, is as old as any of those above, but 
more descriptive. These terms are used reciprocally, i.e., ego 
refers to his wife's sisters' husbands by any of these terms, accord- 
ing to local usage, and they refer to him by precisely the same 
term. These terms are only used in the referential. Vocatively, 
brother terms are usually adopted, or teknonymy is practiced. 

13. I mei W& w y si 

CKi mei 2T and nei mei 28 are alternative and principally de- 

18 Liang Shu 12.4a; Chin shih ts'ui pien 101.26a. The use of nei hsiung ti for 
wife's brothers was confused with that for mother's brother's sons. It may be due to 
the influence of earlier cross-cousin marriage terminology from which the new nomen- 
clature still could not extricate itself. 

"See discussion pp. 194-197. "Shih Ching 12A. 3b; ^gffi. 

*° Nan shih 45. 13b. 23 Han Shu 64A. 10b: £J§ . 

"ffi. 2i Erh Ya: fpjP^fiffO. 

"Luan chen tzu MMP^It' V Ma Yung-ch'ing J^cJftJ. ca. 1110 A. D. 
[1920, Commercial Press ed.] 2. 5b: 3|&, S#. 

26 Ho pi shih lei [cited by Ch'eng wei Lu 7. 17a-b) . 

17 San kuo chih 22. lb. 28 Ibid. 9. 6a. 


scriptive terms. I, as used in the Erh Ya, is interpreted as mean- 
ing the wife's sisters who have married different men, and most 
probably it was originally a man's term. 29 Ti $& is used to mean 
sisters who have married the same man in connection with the 
yin marriage custom, and is more likely to be a woman's term. 30 
Hsiao i is a modern colloquial expression. 

14. I mei fu ^#c^ w y si h 

15. Tang chiu hsiung 'g.MR s of 3 or 5 > w 

16. T'ang chiu sao ®@$t w of 15 

17. T'ang chiu ti @®!& s of 3 or 5 < w 

18. T'ang chiu ti fu ®@i£*f w of 17 

19. T'ang i tzu 1£$&W d of 3 or 5 > w 

20. T'ang i tzu fu ©©#>* h of 19 

21. T'ang i mei @@#c d of 3 or 5 < w 

22. T'ang i mei fu @®#M: h of 21 

Wife's father's sister's children, wife's mother's sister's and 
brother's children are called nei piao hsiung ti for males, and nei 
piao tzu mei for females. It could be further differentiated by 
adding ku, i and chiu, e. g., nei ku piao hsiung for wife's father's 
sister's son older than wife. 

III. Generation of the son 

23. Nei chih P3@ w b s 

Chih was originally a woman's term for brother's child, being 
reciprocal with ku. Since the Chin period it has been used more 
as a man's term for brother's son, hence nei is prefixed, in contra- 
distinction with chih alone. See Table I, 125. In contemporary 
usage, female ego would use chih for brother's son before marriage, 

" Shih Ching SB. 9a; Tso Chuan 8. 24a. 

•• Shih Ching 18D. 5a. Cf. the Shih ming. It seems the term ti cannot be separated 
from the yin marriage custom. Accordingly, when the yin marriage ceased to be 
practiced, ti also ceased to function. 


but after marriage she would use chih for husband's brother's 
son and nei chih for her own brother's son. 

24. Nei chih fu ®®*f w b s w 

25. Nei chih mi ©@ic w b d 

26. Nei chih hsu @®*S w b d h 

27. I wai sheng M#45 w si s 

Ch'i sheng 31 was used circa sixth century A. D. It is more or 
less descriptive, i. e., wife's sheng. 

28. I wai sheng fu ©®^*f w si s w 

29. I wai sheng nii ©®^^C w si d 

30. I wai sheng hsu 003U*§ w si d h 

In certain local usages i chih may be substituted for i wai sheng 
in terms 27-30. Although illogical, it is permitted locally. 

IV. Generation of the son's son 

31. Nei chih sun faMM w b s s 

The Erh Ya uses the term kuei sun, which literally means " re- 
turning grandson." Kuei sun was probably a woman's term, since 
the Erh Ya says that chih's sons are called kuei sun, and chih is 
primarily a woman's term in the Erh Ya. 

32. Nei chih sun fu DDDif w b s s w 

33. Nei chih sun nii DDD^C w b s d 

34. Nei chih sun hsu DDDif w b s d h 

35. I wai sheng sun: wsiss 32 

36. I wai sheng sun fu: w si s s w 

37. I wai sheng sun nii: w si s d 

38. I wai sheng sun hsu: w si s d h 

11 Liang Shu 28. 2a. 

11 For characters, see Nos. 24-30 above. 


In some local usages i chih sun may be substituted for i wax 
sheng sun. Terms 23 to 38 are used by husband and female ego 

Affinal Relatives — Table IV 

Relatives Through Husband 

I. Generation of the husband's father 

1. Kung<&hf 

In the Erh Ya the general term is chiu ; when he is living, chiin 
chiu 1 is used — a reflection of cross-cousin marriage. In modern 
ritual works, the compilers still use chiu for husband's father and 
refuse to employ the modern term kung. If they are afraid lest 
the term be misunderstood, they employ the descriptive nomen- 
clature, e. g., fu of fu JzZX " father of husband " or mu of fu 
^£# " mother of husband/' 

Chang 2 and chung s were used during and somewhat before the 
Han period. Kuan 4 was a local term in the lower Yangtze valley 
circa the end of the T'ang dynasty. All these terms seem to have 
been more or less local, and their degree of prevalence is uncertain. 
The modern term kung dates from about the fourth and fifth 
centuries A. D.; 5 it is also used in the doubled form kung kung. 

Hsien chiu, 6 huang chiu, 7 and hsien tzu 8 are old posthumous 
terms no longer used today. 

2. Foihm 

Ku, and chiin ku only when she is living, are the terms used in 
the Erh Ya — a reflection of cross-cousin marriage. Wei 9 was 
used during the Han period, and wei ku 10 is equivalent to chiin 

x Erh Ya; jg^. 

'Shih Ming; ^. 

* Lii Shih eh'un ch'iu 14.17a: ^. 

*Nan Tang Shu [by Ma Ling] 25.2a: *§£ . 

B Yii t'ai hsin yung chi 1. 16a-21a. 

9 See note 1 . 

7 1 Li 6.1a-2b. 

*Kuo Yu5.9b: jfefr. 

*Shuo Wen; g£. 

10 Wang Nien-Sun [Kuang ya su cheng 6B.5]: j|Ejj!f. 


ku of the Erh Ya. Chia 9 or a chia, 11 are terms used circa the 
fourth and fifth centuries A. D. and surviving quite late as dia- 
lectal forms. M u 12 was also used circa 500 A. D. P'o in ancient 
usage may mean any old woman; its use for husband's mother 
dates from the T'ang period. The use of kung and po for hus- 
band's parents might also be due to teknonymy, since from quite 
early times kung and p'o have commonly been used as grand- 
parents' terms. 

Huang ku and hsien ku are ancient posthumous terms. 

II. Generation of the husband 

3. Pen shen ^# ego, a female 

Ego, a female, might, in speaking to the husband's relatives, 
refer to herself as hsin fu, during the fourth and fifth centuries 
A. D. 13 This custom seems to have been in vogue as late as the 
twelfth century A. D. 14 At present, the proper relationship term 
should be used. 

4. Fu ^c husband 

Chang fu 3b^c and fu hsil are alternative terms. Hsii may be 
used alone for husband. Lao kung 15 and nan jen 16 are colloquial 
rather vulgar, terms. 

Whether or not shift, 17 po, 18 tzu, 19 chun tzu 20 fu tzu 21 and 
chia 22 are actually ancient relationship terms for husband can- 

xx Pei Ch'i Shu 30.4b. Nan shih 33.9a. The chia may be a different rendering 
of ku, as they may be pronounced about the same. Cf. Yen shih chia hsiin 1. 14a. 
12 ££ read 3^C3E*JJ mu. 
ls Shih shuo hsin yu 2B.40b-41a: aE^^^-hES, JrLI^ltif fGR^ % 

ja/hflUWfrtf, TOffM/hlB- Hsin fu literally means "the bride." 

"Shu I 1.12a. 

15 This is used chiefly in southeast China, as in Kiangsu, Kwangtung, etc. It literally 
means " the old male." 

10 A very common term used in the sense of " husband "; literally, " male.'* 

17 Shih ching 3C. 3a: ~±. 

" ffi: see p. 193, note 80. 

19 Shih ching 4C. 2b: ^. 21 Meng Tzu 6A. 4a; Hou Han Shu 113. 9. 

20 Ibid. 7A. 5a. 22 Kuo Yu 6. 8b; Meng Tzu 6A. 6a. 


not be determined. They might be merely general compli- 
mentary terms for man used in the sense of " husband," or simply 
circumlocutory expressions. Tien™ so t'ien 2 * and kao chen 25 
are primarily literary forms ; kao chen is used almost exclusively 
in poetry. 

The wife calls the husband Hang, 26 Hang jen 27 lang 28 and 
ch'ing. 29 Ch'in is a common reciprocal term. All these words are 
old forms, now chiefly retained in literary usage and seldom, if 
ever, used in the vocative. The wife may call the husband by his 
personal name, or just " you/ 5 30 and most prevalently she em- 
ploys teknonymy. 

The depreciatory terms are wai tzu* 1 and cho ju $5^ or yu ju 
S-£. These terms are employed only in refined society. Ordi- 
narily, the wife refers to her husband as t a fife, meaning " he " or 
" him," or by teknonymous and circumlocutory expressions. 

Huang p ( i 32 is an ancient and now obsolete posthumous term. 

5. Po ffi h o b 

H stung kung is used in the Erh Ya. Kung & is often written 
ifc or Mk and is sometimes pronounced chung.™ Hsiung chang 34 
was commonly used during the Han period. Chang is written 4£ 
or **. 

Po means principally father's older brother. Its extension to 
husband's older brother first occurred at about the end of the 
T'ang period. 35 This change can be explained on the basis of 
teknonymy. Po po is more colloquial. 

28 / Li 30. 15b. Tien (heaven) is used in the sense of the " positive " or " male " 

2 * r/f^ is based on the above, employed only in literary usages, i. e., non-vocative 
and non-referential. 

25 -£$5^011* [Yu t'ai hsin yung chi 10.6a]: jf^. 

ae / Id 5. 8a. 

21 Shih ching 6C. 8a; Ming Tzu 8R. llb-12a. 

28 Chin shu 96. 9b. 

28 Shih shuo hsin yu SB. 48b-49a: $$. 

*° Yen shih chia hsiin 1.13b: iiJP^lS, I&M&^o 

Wife calls husband wai and husband calls wife nei\ this practice dates from the 
middle of the first millennium A. D. Cf. Heng Yen Lu 3.11a. 

32 Li Chi 5.22a: j^. 83 Erh Ya commentary. 8 * Shih Ming: ft^. 

35 See p. 193, note 81. 


6. Mu mu fiH* hobw 

Ssu fu is the ancient term used in the Erh Fa, 86 and is scarcely 
known today. Mm mu first came into use during the Sung period. 37 
It is sometimes written *Wt ; the pronunciation differs slightly in 
different localities. 

7. Shu &hyb 

Hsiao shu,** shu lang,™ and hsiao lang 40 are alternative terms 
that date from the fourth and fifth centuries A. D. Hsiao shu is 
more colloquial, shu lang and hsiao lang are more literary. Shu 
is used in the &rh Ya. 

In the lower strata of present-day society, the younger brother- 
in-law can usually " play jokes " with the older sister-in-law. This 
circumstance is primarily based on the popular assumption that 
the younger brother-in-law is always a minor and that the older 
sister-in-law assumes a kind of maternal attitude. 

8. Shen shen *$*& h y b w 

Ti fu is an ancient term. 41 Ssu and ti are also used in the Erh 
Ya to mean sisters who married the same man ; the younger calls 
the older ssu, and the older calls the younger ti.* 2 This usage is 
probably connected with the yin marriage. When used for hus- 
band's brother's wives, the fu should be appended as it is in the 
Erh Ya and / Li (33. 2a-b) . 

Shen shen was first used during the Sung period. 43 Shen was 
originally a term for father's younger brother's wife, and its 
extension to husband's younger brother's wife is certainly 

Female ego and husband's brothers' wives may refer to each 
other as ti ssu, as recorded in the Erh Ya, and as hsien hou 44 and 
chou li, A * as used during the Han period. Ti ssu and hsien hou 

■• mm- 

87 See p. 201, note 104. 

88 Skik Chi 69. 15b-16a. 

89 Wen Hsiian 40. 6b. " #H> JH&- 

*• Chin Shu 96. 9b. " See p. 201, notes 103^ 104. 

" Cf. Erh ya. " Han Shu 25A. 18b: jtffc. 

"Fang Yen 12.1a: Hffi. Chu is a synonym of 4ft. 


are now obsolete. Chu li is the prevailing term at present ; it is 
reciprocal and used only in the referential. 

9. Ku *£ h si 

Ta ku izik may be used for the older sister, and hsiao ku /MS 
for the younger sister, of the husband. On the other hand, ta ku 
need not necessarily indicate that the sister is actually older than 
the husband, since the female siblings may be counted in a sepa- 
rate numerical series. Hence ta and hsiao may only indicate the 
seniority and juniority among the husband's female siblings. In 
fact, ta ku and hsiao ku may be both younger than the husband. 
These terms are mainly used for the husband's unmarried sisters, 
but they may continue to be used after their marriage, though 

In the Erh Ya 9 husband's older sister is called nil kung :&&> 
and the younger, nil mei icfa. Shu mei was used during the Han 
period. 46 Hsiao ku was first applied ca. the fifth century A. D. 
Since ku originally meant father's sister, its extension is attri- 
butable to teknonymy. 

10. Ku fu **:£ h si h 

Ku fu also means father's sister's husband, and its application 
to husband's sister's husband is undoubtedly teknonymous. 

'Cf. Hou Han Shu 114.8. 


Having discussed the system in its morphological and historical 
aspects, I shall here venture upon a few concluding remarks. It 
may safely be asserted that during the last two thousand years 
the system has undergone a series of changes both in its structural 
principles and terminological categories, yet has retained many 
features of the old system. The latter stability seems to be re- 
lated to the continuity of Chinese civilization as a whole. As 
regards the changes, there is, generally speaking, a broad historical 
correlation with the changes in the development of Chinese society. 
One notes that practically all the kinship changes occurred during 
the last two centuries of the first millennium B. C. and the whole 
of the first millennium A. D. During this period the system was 
in a state of flux. Many old terms were dropped, changed, or 
delimited in connotation. New terms were introduced, as if by 
way of experiment; some were incorporated into the system, 
others fell into oblivion. Almost all the new terms used in the 
modern system originated at this time. The whole system was 
finally stabilized during the T'ang period, after a thousand years 
of constant transformation and confusion. 

This millennium likewise was a period of civil and social strife, 
the aftermath of the dissolution of the old feudal system. To be 
sure, the entire social structure was not suddenly transformed; 
many of the old social institutions lingered on, though in slightly 
modified forms. 1 Nevertheless, the evolution of the new social 
order was begun. It was a slow and gigantic process, accompanied 
by periods of alternating political and social tranquillity and chaos, 
of reactionary and progressive thought. This was an age of widen- 
ing contacts with outside influences, especially from the third to 
the sixth centuries A. D., the " Dark Ages " of Chinese history, 
when nearly all the territory north of the Yangtze was overrun 
by less civilized peoples from the northern steppes ; the resultant 

1 There are students who would even consider present China a fundamentally 
feudal society. This is somewhat an exaggeration and depends on one's definition of 




large waves of Chinese migration southward resulted in the 
efflorescence of the shih tsu tfr^ organization and its excrescent 
manifestation in the men /a P!ffi 2 system of official recruitment. 
The entire process is too complex a subject to be dealt with here, 
but it shows a general chronological correlation with the develop- 
ment of the kinship system. When the kinship system crystalized 
toward the end of the first millennium A. D., Chinese society still 
continued to evolve. The kinship system, being a more conserva- 
tive institution and in some ways a stabilizing mechanism for 
other social institutions, has remained essentially the same as it 
was during the Tang period. 

The individual parts of the system have exhibited varying 
amounts and rates of change, i.e., the changes have been dif- 
ferential. The nomenclature for sib relatives has experienced 
relatively little alteration, although there have been refinements 
in the degrees of differentiation and, for some terms, changed 
connotations. This may be due to the fact that, although the 
old sib system, tsung fa, has been transformed into the modern 
sib organization, shih tsu, the sib principle has remained the basis 
of kinship evolution. The increased descriptive efficacy repre- 
sents merely refinements of superficial features of the system, 
correlated with the elaboration of the mourning ritualism. 

Most remarkable changes have occurred in the terminology for 
non-sib relatives, especially in the nomenclature for affinal rela- 
tives, which Aginsky calls "basic terminology." 3 A glance at 
Tables III and IV will show how radical and complete the changes 
are. It is a generally accepted fact among social anthropologists 
that the affinal terminology is extremely sensitive to variations 
in marital relationships. But have the Chinese marriage regula- 
tions radically changed during this period? This question finds a 
simple answer in history. The most important factor in the regula- 
tion of marriage in China, from the first millennium B. C. down 
to the present, has been sib-exogamy supplemented by the genera- 

2 Men fa, as a system of official recruitment, is based on sib connections. Cf. *H3S 

3 B. W. Aginsky: Kinship systems and the forms of marriage, Memoirs, American 
Anthropological Association, 45 (1935), 14. 



tion principle. It has been pointed out above that as time went on 
a gradual stiffening of this rule took place, but there was little 
actual change. In general, the effect of marriage regulations upon 
the kinship system has been so small that we are justified in 
ignoring it. We can also, in the manner employed before, dispose 
of the sororate and the levirate as moulding influences on the 

Cross-cousin marriage, however, presents a very different prob- 
lem, for it is upon this that the affinal terminology of the old 
system was undoubtedly based. If we assume that the decline in 
the frequency of this form of marriage resulted in the breakdown 
of the old affinal nomenclature, we must still explain the origin 
of the new terminology. These new terms are, in my opinion, not 
the products of new forms of marriage, but are the result of the 
operation of teknonymy. Marital relationships as a whole have 
had little influence upon the modern system. The reason for this 
may he in the fact that the Chinese marriage regulations are 
purely restrictive, not prescriptive, i.e., aside from certain re- 
strictions connected with sib-exogamy and generation, there is 
complete freedom of choice. 

The morphological configuration of the Chinese sytem has 
puzzled many a student. Morgan, in generalizing on the system, 
vacillated between his Malayan and Turanian, saying that " it 
falls below the highest type of the Turanian form, and affiliates 
wherever it diverges with the Malayan." 4 Lowie, apparently 
using the same material, considered the Chinese system either a 
" generation " or a " bifurcate merging " system 5 — which is equiva- 
lent to saying that it is either Malayan or Turanian. T. S. Chen 
and J. K. Shryock, using Lowie's system of classification, call the 
Chinese system " bifurcate collateral." 6 Kroeber is of the same 
opinion, but says, " the Chinese system appears to consist of a 
* classificatory,' that is non-descriptive, base, which has been made 
over by additions into a ' descriptive ' system similar in its work- 
ing to the English one, in fact is more precisely and successfully 

* Morgan: System . . . , op. cit., 413. 

* R. H. Lowie: Relationship Terms, Encyclopaedia Britannica 1 *. 
6 Op. cit., 6*7. 


descriptive than this." 7 Compare this with Morgan's remark that 
the Chinese system " has accomplished the difficult task of main- 
taining a principle of classification which confronts the natural 
distinctions in the relationships of consanguinei, and, at the same 
time, of separating those relationships from each other in a precise 
and definite manner." 8 

Actually, the Chinese system is not one that lends itself to any 
simple characterization in terms such as " classificatory " or 
" descriptive." It must first be understood in the light of its own 
morphological principles and historical development. In Morgan's 
definition of the terms, the Chinese system is both classificatory 
and descriptive. This is not an " inconsistency " from the point 
of view of Morgan's system, but a characteristic of a system 
moulded by diverse factors of a counteracting nature. It is the 
elucidation of these underlying factors that is of scientific import 
rather than any particular characterization. This problem has 
been approached through a detailed analysis of the changes which 
the Chinese system has undergone during the documentary 
period of its history. Insofar as the data have permitted, corre- 
lative sociological facts have been evaluated with reference to 
these changes and the nature of possible dynamic factors indicated. 

7 Process in the Chinese Kinship, op. cit, t 151. 

8 Morgan: Systems . . . , 413. 


The following list has been prepared for the sole purpose of avoiding the constant 
repetition of bibliographical information in the text. The works are arranged alpha- 
betically according to the romanized brief title as used in the text and notes. Since 
full bibliographical information is usually given in the notes for Chinese works cited 
only once or twice and for European works referred to, they are not given here. 

Chan kuo ts'e HcSSj $*$:&: commentaries by Pao Piao ffiM , circa 

1150 A.D., and Wu Shih-tao &M9L, 1283-1344 A.D.; 

Ssu pu ts'ung k'an EMfU. 
Ch'ang li chi 13£ft£&: Collected works of Han Yii *>*, 768- 

824 A.D.; Ssu pu pei yao HflSWHS. 
Ch'ao ching ch'ao wen chi M:fflL$k%.M.: Collected works of Cheng 

Chen %&, 1806-1864 A.D.; J(!ftt**MF edition. 
Ch'eng wei lu ffiffltt: by Liang Chang-chu M*&, 1775-1849 

A.D.; 1875 edition. 
Cheng tzu t'ung JE^il: compiled circa 1670 A. D. by Liao Wen- 

ying &%.%; 1670 edition. 
Chi yiin ftiB: compiled by Ting Tu TSL, 990-1053 A. D., and 

others ; Ssu pu pei yao edition. 
Chiang Wen-t'ung chi tlj&Ilfc: Collected works of Chiang Yen 

CCfS, 444-505 A. D.; Ssu pu ts'ung k'an edition. 
Ch'ien yen fang chin shih wen pa wei *BW£&#3fcttfc: by 

Ch'ien Ta-hsin «*IW, 1727-1804 A.D.; 1884 »W£&# 

Chin shih li ttW: by P'an Ang-hsiao flIIMf, circa 1300 A. D.; 

®&mmmt edition. 

Chin shih ts'ui pien &ft%1&: compiled by Wang Ch'ang 3E&. 

1727-1806 A.D.; J&PI& edition. 
Chin shih yao li &55W: by Huang Tsung-hsi %&M, 1609- 

1695 A.D.; imH \UB *# edition. 
Chin shu W#: by Fang Chiao BWi, circa 630 A. D., and others; 

1894, fflJ&rM edition. 
Ch'in shu chi MEtl: by Cheng Chen %% % 1806-1864 A.D.; 

HWS* edition. 


Chiu Tang shu *Jf#: by Liu Hsu SW, 887-946 A. D., and 

others; 1894, MXW^ edition. 
Ch'u-chiang wen chi ffiff^tft: Collected works of Chang Chiu- 

ling SBAtfr, 673-740 A. D.; Ssu pu tsung k'an edition. 
Erh ya SMfSJtfifc: Commentary by Ho I-hsing Ms&Al , 1757-1825 

A. D.; Ssu pu pei yao edition. Since the section referred 

to is the short Shih Cliin #H, Relationship Terms, no 

page reference is given in the notes. 
Fang yen 2fW: by Yang Hsiung tM£, 53 B. C.-18 A.D.; com- 
mentary by Kuo Fo fB*I, 276-324 A.D.; Ssu pu tsung 

Jc'an edition. 
Han fei tzu St^P: by Han Fei St*, ?-324 B. C. ; Ssu pu pei yao 

Han shu «#: by Pan Ku BE@, 32-92 A.D.; 1894, MXWJ& 

Heng yen lu U §£i: by Ch'ien Ta-hsin $&JzWi 9 1727-1804 A. D.; 

1884, Ifff^rt edition. 
Hou Han shu &M^ : by Fan Yeh ?E8£, ?-445 A. D.; 1894, f^I^C 

^M edition. 
Hsiao ching #S&ifc: 115, K*J+H«aBt edition. 
Hsin shu ^Pflfr: by Chia I HM, circa second century B. C.j Ssu 

pu pei yao edition. 
Hsin Tang shu #fJi»: by Ou-y ang Hsiu RHtf, 1007-1072 A. D., 

and Sung Ch'i M5IB, 998-1061 A. D.; 1894, f*J2fc*^ 

Hsiieh lin &W; by Wang Kuan-kuo 3EftH, circa 1140 A.D.; 

S^JtWK edition. 
Hua yang chi #fi§H: by Ku K'uang Hffil, circa 8th and 9th 

centuries A.D.; 1855 tfi$£ edition. 
Huai nan tzu fiPfr?: attributed to Liu An *J£, P-122 B.C.; 

1876 fflULWm edition. 
I ching M&&&: 1815 IftjW-hHjK&SS edition. 
I li M&Bft: 1815 BcSlJ+HJK^etE edition. 
Jih chih lu Bailiff: by Ku Yen-wu WjfcS, 1612-1681 A. D., 

commentary by Huany Ju-ch'eng ^^C^; 1872 8IMb#2BC 

<FJ§i edition. 


Jung chai sui pi &M^E: by Hung Mai $F*S, 1123-1202 A. D.; 

Ssu pu ts'ung k'an SflS edition. 
Kai yii ts'ung k'ao &&£#: by Chao I MM, 1727-1814 A.D.; 

1790, ##£tttt£# edition. 
Kuang shih ch'in ^P§i: by Chang Shen-i 36»fll, based on 

mmm&mm. #ss# edition. 

Kuang ya W «M5ft»: by Chang I 3SS, circa 230 A. D., com- 
mentary by Wang Nien-sun 3E&», 1744-1832 A. D.; 1879, 
?i^#^ edition. Since the section referred to is the Shih 
Ch'in SS, Relationship Terms, 6. 1.-6, no page reference 
is given in the notes. 

Kuang yiin 0HB: revised by Ch'en Feng-nien MS*, 961-1017 
A. D., and others; Ssu pu ts'ung k'an edition. 

Kung-yang chuan &3£&8t: 1815, KlM+Htf ^K edition. 

K'ung ts'ung tzu JLH^: attributed to K'ung Fu HM, circa 200 
B.C.; Ssu pu ts'ung k'an edition. 

Kuo yii Mm: commentary by Wei Chao ^Bg, 204-273 A.D.; 
Ssu pu ts'ung k'an edition. 

Li chi ma^tt: 1815 giGfffl-hSK&ift edition. 

Liang shu ^#: by Yao Ssu-lien jftSft, ?-637 A.D.; 1894 MX 
HF/pj edition. 

Liu Pin-k'o wen chi £95fc*3t£: Collected works of Liu Yu-hsi 
$i$M , 772-842 A. D. ; Ssu pu pei yao edition. 

Lu Shih-lung wen chi W±ntXM: Collected works of Lu Yiin 
PlU , 262-305 A. D. ; Ssu pu ts'ung k'an edition. 

Lun yii fM&&8fc: 1815, RcSJ+HJK&iBE edition. 

Lu shih ch'un ch'iu SK#I^: attributed to Lii Pu-wei B^$- 
?-235 B. C; Ssu pu pei yao edition. 

Meng ch'i pi fan £8S*5fc : by Shen Kua J5M8, 1030-1094 A. D. ; 
&j8SW edition. 

Meng tzu SHP&Bfc: 1815 KilJ+HtS&Hit edition. 

Ming lii chi chieh fu li WfaMkMffl&l: compilation of 1585 A.D., 
KS+H* ; 1908, «£?*#» edition. 

Ming tao tsa chih WJMfcfe: by Chang Lei »*, 1052-1112 A.D.; 
M8:XM 'baft edition. 

Mo chi RiS: by Wang Chih If 5, circa 1120 A. D. ; 1918 «#* 


Nan shih ftfjll: by Li Yen-shou ^5£#, circa seventh century 

A.D.; 1894 MXWM edition. 
Nan Tang shu ^JiS#: by Ma Ling H^ circa 1100 A. D.; Ssu 

pu ts'ung k'an $RH edition. 
Neng kai chai man lu l&Qfcffi&tt: by Wu Tseng ^#, circa 1150 

A.D.; $&&&&!& edition. 
Pa ch'iung shih chin shih pu cheng ASfe^E&^JW JE: compiled by 

Lu Tseng-hsiang &#**, circa 1850 A. D. ; ^H£']R#~&*i 

Pai ching fang wen chi ^FJKfiXH: by Tsang Yung tiUtf, 1767- 

1811 A.D.;1920 _L7C^ROT edition. 
Pai hu t'ung &ifeatefcS: attributed to Pan Ku M, 32-92 A. D., 

and others; commentary by Ch'en Li $.+£, 1809-1869 

A.D.; 1875 Mffi^M edition. 
Pei Ch'i shu *#*: by Li Pai-yao &W& $ 565-648 A.D.; 1894 

MX^Im edition. 
Pei meng so yen ^fci^iflf : by Sun Kuang-hsien ZfcftM, ?-968 

A.D.; M&S* edition. 
Pei shih it A: by Li Yen-shou ^5I#, circa 7th century A.D.; 

1894 MX^ffi edition. 
Fieh chi ^IS; by Liang Yii-sheng *3£*1, circa 1780 A.D.; ffi 

&dr^ edition. 
San kuo chih HB* : by Ch'en Shou Bill, 233-297 A.D.; 1894 

|pJ2St"#jR7 edition. 
Shih ch'ang t'an ffSSfe: anonymous; circa 1100 A.D.; "S"JM*tS 

Shih chi ftffi: by Ssu-ma Ch'ien *HS», P145-74? B.C.; 1894 

WtX&ffi edition. 
Shih ching ^I3$fe5fc: Ssw pw pei yao edition. 
Shih ming fl^itS: by Liu Hsi 91 KB, circa 200 A. D„ com- 
mentary by Pi Yuan *5c, 1730-1797 A.D.; JSfflifc* 

edition. The part referred to is the Section II: Shih Ch'in 

Shu ffllJil, Relationship Terms. No page reference is 

given in the notes. 
Shih shuo hsin yii ifrtft$fi&: by Liu I-ch'ing ftl&K, 403-444 

A, D., commentary by Liu Hsiao-piao §!I#1^, circa 530 

A. D.; Ssu pu ts'ung k € an edition. 


Shu ching 1ft#&5l : 1815, K IW+Hff ffigt edition. 

Shu i »«:by Ssii-ma Kuang lUfft, 1019-1086 A.D.; &Wtt 

i§C edition. 
Shu p'o BM: by Tai Shih MM, circa 1220 A.D.; MWM 

Shuo wen tfcXfl?^: by Hsu Shen fHi, ct'rca 200 A. D., com- 
mentary by Hsii Hsiian %&, 916-991 A. D.; Ssu pu ts'ung 

k'an edition. 
Sung shu *#: by Shen Yo fcfctt, 441-513 A.D.; 1894 IpI**^ 

Ta Tai li chi jzBMXB,: by Tai Te M, circa 100 B.C.; Ssu 

pu ts'ung k'an edition. 
T'ang chih yen M.W : by Wang Ting-pao Mffic, 870-954 (?) 

A.D.; M^«# edition. 
T'ang lu su i tkMWBM : codified by Ch'ang-sun Wu-chi $M 

Mfe, ?-659 A.D. and others. Ssu pu ts'ung k'an Hfil 

Ts'ai Chung-lang chi H+fl|5*: Collected works of Ts'ai Yung 

Hii, 133-192 A. D.; Ssu pu ts'ung k'an edition. 
Ts'ao Tzii-chien chi W^jtlfc: Collected works of Ts'ao Chih 

Wffit, 192-232 A. D.; Ssu pu ts'ung k'an edition. 
Tso chuan &&&B: 1815, KSU+HIKaa edition. 
Tsung fa hsiao chi SBM4B: by Ch'eng Yao-t'ien ^&E9 , 1725- 

1814 A. D. ; ftl&M edition. 
T'ung su p'ien ftftft: by Chai Hao Sffl, P-1788 A.D.; WPS. 

3§f edition. 
T'ung tien MM: compiled by Tu Yu ttftf, 735-812 A. D.; 1896, 

fflUX&JH] edition. 
T'ung ya &*ft byFangl-chih l£JH£, circa 1650 A.D.; Afttt 

Wang Shih-chung chi If^H: Collected works of Wang Ts'an 

3E^, 177-217 A.D. ; tftStfcttWH^SMS edition. 
Wei Wu-ti chi ftlft**: Collected works of Ts'ao Ts'ao Wt&- 

155-220 A. D.; StfS^J&TS'Hig^fe edition. 
Wen hsiian XUti&LM: compiled by Hsiao T'ung jf tt, 501-531 

531 A. D., commentary by Li Shan $#, ?-688 A. D. and 

others. Ssu pu ts'ung k'an edition. 


Yen shih chia hsiin fiK^IW: by Yen Chih-t'ui M±M, 531-591 

A. D. ; Ssu pu pel yao edition. 
Yen tzu ch'un ch'iu S^^^C: attributed to Yen Ying f?H> 

P-493 B. C. ; S^ pu pet yao edition. 
Yin hua lu Hf£3t: by Chao Lin ffi*, circa 840 A.D.; WM 

Yii p'ien 3£lf : compiled in 543 A. D. by Ku Yeh-wang HIM 

augmented in 760 A.D. by Sun Ch'iang $&$&, revised in 

1008 A. D. by Ch'en Feng-nien WM¥, and others; Ssu pu 

ts'ung Van edition. 
Yii t'ai hsin yung chi ~^MM^%\ compiled by Hsu Ling !&!& 

507-583 A. D.; Ssu pu ts'ung lean edition. 
Yuan shih chang-ch'ing chi TCRJIiiE^iE: Collected works of Yuan 

Chen TCfM, 779-831 A. D.; Ssu pu tsung k ( an edition.